Professional Documents
Culture Documents
Das Pal Volume 2
Das Pal Volume 2
ENGINEERING
MATHEMATICS
Volume-IIA for 2~ ,0 er
(For CSE & In
"
f~' 1t ~~
~l
6',?'~
~ ,
I~'O~//J}f!~ ~
B. K. PAL, M.Sc.,Ph.D. (\~\ _ ~.:
(P. N. Banerjee Gold Medalist & Recipient of ;t~,4tGe.1f6~~)
Head of the Department of Mathematics(Ret. . ,-"
Maulana Azad College, Kolkata
Formerly, Reader in Mathematics
Kalyani Government Engineering College, Nadia, West Bengal.
K. DAB, M.Sc.,B.Ed.,Ph.D.
Associate Professor and Head of the Department of Mathematics,
Krishnanagor Government Colleg"e, Nadia, West Bengal
Formerly,
Associate Professor, A.B.N. Seal College, Coochbehar, West Bengal
Assistant Professor and Head 'of the Department of Mathematics
Kalyani Government 1'__ : : __ 1'_11 ~Tadia, West Bengal.
Acadtmy or Technology
1111111111111111111111
49296
ISBN-97S-93-S0673-16-5
Module No. 1.
Basic Probability:
Probability spaces, conditional probability, independence;
Discrete random variables, Independent random variables, the Multinomial
distribution, Poisson approximation to the Binomial distribution, infinite
sequences of Bernoulli trials, sums of independent random variables; Expectation
of Discrete Random Variables, Moments, Variance of a sum, Correlation
coefficient, Chebyshev's Inequality. 11L
Module No.2
Continuous Probability Distributions:
Continuous random variables and their properties, Distribution functions and
densities, Normal, Exponential and Gamma densities 4 L
Module No.3
Bivariate Distributions:
Bivariate distributions and their properties, distribution of sums and quotients,
Module No 4
Basic Statistics:
Measures of Central tendency, Moments, Skewness and Kurtosis, Probability
distributions: Binomial, Poisson and Normal and evaluation of statistical
parameters for these three distributions, Correlation and regression - Rank
correlation. 8L
Module No 5
Applied Statistics:
Curve fitting by the method of least squares- fitting of straight lines, second
degree parabolas and more general curves. Test of significance: Large sample
Module No 6
Small samples:
Test for single mean, difference of means and correlation coefficients, test for
attributes. 4L
CONTENTS
MODULE-I
m;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;_B_AS~IC~P_R_O_B_A_B_IL_IT_Y;;;;;;;;;;;;T;;;;H;;E;;O;;;;R;;;;Y
1.1. Introduction 1
1.2. Introductory Terminologies 1
1.3. Classical Definition of Probability. 3
1.4. Theorems on Probability. 4
1.5. Axiomatic definition of probability. 7
1.6. Conditional Probability. 10
1.7. Independent Events 11
1.8. Baye's Theorem 12
1.9. Illustrative Examples. 12
Exercises 26
Short Answer Questions 26
Answers 30
II Long Answer Questions 31
Answers 37
III Multiple Choice Questions 37
•
m Answers
m
fJl
~~~~~~~~
3.1. Random Variable.
3.2. Probability Mass Function and Discrete Distribution
DISCRETE RANDOM VARIABLE
AND ITS EXPECTATION
62
63
3.3. Distribution Function or Cumulative Distribution Function
(For Continuous & Discrete) 64
3.4. Expectation or Mean of a Discrete Random Variable 67
3.5. Variance and Standard Deviation ofa Random Variable 69
x ENGINEERING MATHEMATICS -IIA
[!j]1;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;S;;;;PE;;;;C;;;;IAL;;;;;;;;;;;;;;;TY;;;;;;;;;;;P;;;;E;;;;O;;;;F;;;;D;;;;;;;;;;;IS;;;;C;;;;RE;;;;T;;;;E;;;;;;;;;;;D;;;;IS;;;;T;;;;RI;;;;;;;;;;;BU;;;;T;;;;I;;;;O;;;;N
4.1. Introduction 94
4.2. Binomial Distribution 95
4.4. Poisson Approximation to Binomial Distribution. 112
Exercises 117
Short Answer Questions 117
Answers 119
II Long Answer Questions 119
Answers 124
III Multiple Choice Questions 125
Answers 127
m
MODULE-II
ml;;;;";;;;;;;;;;;;;;;;;;;;;;;;;;;S;;P;;E;;C;;IAL;;;;;;;;;;;TY;;;;;;;;;P;;E;;O;;F;;C;;O;;N;;T;;IN;;;;;;;;;U;;O;;U;;S ;;D;;I;;ST;;R;;I;;B;;UT;;I;;O_N
MODULE-ill
rn1 BIVARIATE DISTRIBUTIONS
m~;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;!;;;( C;;;;O;;;;N;;;;T;;;;I;;;;NU;;;;O;;;;U;;;;S;;;;V;;;;A;;;;R;;;;I;;;;AT;;;;E~S
)
MODULE-IV
~~;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;B;;;;A;;;;S;;;;IC;;;;;;;;;;;S;;;;T;;;;AT;;;;I;;;;ST;;;;I~CS
~101~
~
C_O_mffi ~_T_IO_N_,~R_E_G_RE
RANK CORRELATION
__ SS_IO~N
MODULE-V
lrn~
11.1 Introduction
C_UR_V_E_FI
__ TT_IN_G
__ BY_ME T_HO_D_O_F_L_E_AST S_Q_UA_RE;;;;;'
412
11.2. Least Square Regression Curve 413
11.3. Normal Equations 414
11.4. Finding best fitted straight lines 415
11.5. Finding best fitted straight line when (X, Y) assumes n pair of
datas exl'yl)'(xz'yz)······(xn'yn) 415
11.6. Finding best fitted Second Degree Parabola when (X, Y)
assumes n pair ofdatas (Xpyl)'(XZ'yz),-·····(xll,yll) 421
xiv ENGINEERING MATHEMATICS-IlA
MODULE-VI
D SMALL SAMPLE TEST OF SIGNIFICANCE
li]]iiiiiiiiOijiiiiiiiiOijiiiiiiiiOijiiiiiiiiOijiiiiiiiiOijiiiiiiiiOij;;;B;;;A;;;S;;;ICiiiiiiiiOijP;;;R;;;O;;;B;;;AB;;;I;;;L;;;IT;;;YiiiiiiiiOijT;;;H;;;E;;;O;;;R;;;;;Y
For example, if A = {TH, HT} where S = {HH, IT, TH, ~T}, then
1= {HH, IT}.
ote. S = <I> ; ~ = S ; (1) = A
Simultaneous Occurence of two Events.
Let Al and A2 be two events. Then the set Al (\ A2 represents the
simultaneous occurence of the two events Al and A2• This event is also
denoted by AI A2 .
For example, in the experiment 'rolling a die' let Al = 'Even face'
A2 = 'Multiple of three'. Then Al (\ A2 = {6} is the event whose
occurence shows the simultaneous occurence of Al and A2.
At least one of Two Events.
Let Al and A2 be two events. Then the set Al v A2 represents 'at
least one of Al and A2 '. This event is also denoted by Al + A2 .
For example, in the experiment rolling a die let Al = Even face
={2,4,6},A2= Multiple of three ={3,6}. Then A\vA2 ={2,4,6,3}
is the event whose occurence shows the occurence of at least one of
'e en face' and 'Multiple of 3'.
isjoint or Mutually Exclusive (m.e) Events. If two events AI' A2
have no common sample points i.e. if Al (\ A2 = <1>, they are called
Mutually Exclusive Events.
For example, in a previous example if Al ={HH, IT}
A2 = {HT, m}, then Al n A2 = <I> . So, Al and A2 are mutually exclusive
events. Two me events cannot occur simultaneously.
BASIC PROBABILITY THEORY 3
airwise Disjoint Events. Let AI' A2, ... , All be n number of events.
Events Aj(i = 1,2,···, n) are said to be pairwise disjoint if no two of them
have any common event points i.e. if Aj (\ A j = <1>, i * j and i, j = 1,2,· .., n .
Exhaustive Events. Two or more events are said to be exhaustive if
at least one of them neccessarily occurs or in other words the events
AI> A21 are exhaustive if AI U A2 U A3 u··· = s.
For example, in the experiment of throwing two coins once, the events
AI = {HH}, A2 = {IT} and A3 = {HT, TH} are exhaustive.
ementary or Simple Event. An event containing exactly one sample
point is called Elementary Event. .
n(A)
P(A) = n(S)" ... (1)
Pro . mitted.
heorem 8. For any event A. p(l) = 1- P(A) where 1 is the
complementary event of A.
Proof: We have A and 1 are mutually exclusive events and
A u 1= S , the certain event.
:. P(AU 1) = p(S)
or, p( A) + p( 1)= 1, since A and 1 are mutually exclusive.
:. p( 1) = 1- p( A) .
Theoem 9. If AI' A2' ... ,All are mutually exclusive and exhaustive
n
events, then L p( A; ) = 1 .
;=1
n 11
Proof:
n(AI) n(A2)
(i) Al C A2 =>
n(AI) $n(A2) =>-- $--=> P(AI) $P(A2)
n n
BASIC PROBABILITY THEORY 7
Then {H}u{T}={H,T}
1 2
:. P({H}u{T}) = P({H,T}) = 1= 3+ 3 = P({H}) + P({T}),
i.e., Axiom ill is satisfied.
Thus this function P represents probability. In particular we can say
probability of Head is j.
AU the Theorems which were deduced for the Frequency definition
of probability can be deduced for Axiomatic definition also. Some ofthems
are shown in the next page.
8 ENGINEERING MATHEMATICS-lIA
=0
Hi) For any two events AI, Az c S and AI c Az .
11 n
i.e., P(UA;) ~ L:P(A;)
;=1 ;=1
(2)
:. ~9tA;)~ ;~t(A;)
;=1
:. ~:stA;)~:~tt P(A;).
Thus the theorem holds for n = m + 1 whenever it is hold for n = m .
But the theorem holds for n = 1,2 . Hence by induction, the theorem is
true for any positive integral values of n.
For any events AI> A2, ... , An,
n
p( AI (\ A2 n ...(\ An ) ~ 1- L p( AI)
;=1
n
or, 1- P(AI r, A2 (\ .. ·n All) ~ LP(A;)
;=1
11
.. p( AI r, A2 (\ ... r, A" ) ~ 1- L p( A; )
;=1
1
10 ENGINEERING MATHEMATICS -IIA
. Conditional Probability.
We consider two events A and B connected with a random experiment
E. Let us make the hypothesis that the event A has occured n( A) times
and B occurs simultaneously with A n( An B) times in the n repetations
n(AnB)
of experiment E .. Then ratio n( A) is called the conditional
probability of B on the hypothesis that A has already occured and is
denoted by p(~). .
p(B/) =n(AnB)
/A n(A)
n(AnB)/n P(AnB)
= n(A)/n = p( A) ,provided p( A);t 0
Similarly the conditional probability of A on the hypothesis that B has
... P ( An/:O:Ai)
provided p(n Ai);t 0, i = 1,2"", n-l
Proof: Left as an exercise.
BASIC PROBABILITY THEORY 11
l{A;)l{~~)
Ii) 1{A;/~) l{A) ,for i=1,2,"',n [W.E.UTech.2008]
1 .
(b) (i) No; because p(AB) =4~ 0, i.e., AnB ~ c/>
7
(ii) No ; because P( A u B) = 12~ 1, i.e., Au B ~ S
1 1
(iii) No; because P( A) = 2' P(B) = '3 i.e., P(A) ~ P(B)
(iv) No; because P(AB)~ P(A)·P(B), here
1 1
P(AB) ='4 but P(A).P(B)=6'
Example 4. In an examination 30% of the students failed in Physics, 25%
in Mathematics and 12% in both Physics and Mathematics. A student is
selected at random' Find the probability that (i) the student has failed in
Physics, if it is known that he has failed in Mathematics.
(ii) the student has failed at least one of the two subjects
(iii) the student has passed at least one of the two subjects.
(iv) the student has passed in Mathematics if he failed in Physics.
Let A and B denote the events" a student failed in Physics" and "a
student failed in Mathematics" respectively. Then
P{A)=0·30, P(B)=0·25, P(ArlB)=0.12
Now (i) probability that a student has failed in Physics if it is known
that he has failed in Mathematics is
- P(AB) 0·12
P(B/A)=I-P{B/A) =1--- =1--=0.60.
P(A) 0·30
BASIC PROBABILITY THEORY 15
P(A )-~x~- 20
2 - 14 14 - 196
7 9 63
P(A3)=14X14= 196
2 20 63 85
So the required probability is 196 + 196+ 196 = 196 .
Example 6. Two cards are drawn from a well-shuffled pack. Find the
probability that at least one of them is diamond.
Let A be the event that at least one of the drawn cards is diamond.
Then A be the event that the drawn cards is not diamond.
. -) _ 39C2 _.!...-9
A
P() 19 15
:. P(A)= 1- A = 1--=-
.. P ( - 52~ - 34 34 34
15
So, the required probability is 34.
=p(SIn1"Ol)=P{(SU10n =1-P(Su10)
= 1- {P(S) + P(IO)- p(S n IOn
=1- C~ +~ -
1
5 2)= 1- ~ = ~3 "
Example 8. Two dice are thrown n times in succession. What is the
probability of obtaining double six at least once.Hence find the minimum
number of throws so that the probability of obtaining double six' at least
or, (35)"
- > -1 or, nlog (35)
- > -log2
36 2 36
log2
or, n> ::::24.6 :. n ~ 25
log36 -1og35
Hence the minimum number of throws is 25.
xample 9. Show that the probability of occurrence of only one of the
events A and B is p( A) + p( B) - 2P( AB) . [ WB. U. Tech 2006]
Let C be the event of occurence of only one of the events A and B.
BASIC PROBABILITY THEORY 17
Then
C=(AuB)-(AnB) s
=(A-B)u(B-A) A B·
:. p(C)=P(A-B)+P(B-A) ... (1)
[ .: A - Band B - A are disjoint]
Now A=(A-B)+AB
~
'-t-i!
:. P(A) = P(A -B)+ P(AB)
:. P(A-B)=P(A)-P(AB) ... (2) 'CADE~V n~ Tw:r.HNOL~
Again B = (B - A) + AB l_.f:'- .' ~ I.. ~ Q / ~ V,.0 ~ -I ~
.: P(B)=P(B-A)+P(AB)
•.. ~~:L~.b.Oatc._._ ..••••
.: P(B- A) = P(B)- P(AB) ... (3)
J
ENGINEERING MATHEMATlCS-IIA
18
n-k k
Then the probability for both the drawn cards are of black is --.-
n n
k n-k
and that of red is -. -- .
n n
So the probability for both the drawn cards are of the same colour IS
n - k . k +!!... n - k = 2 (n - k )k k = 1 2 ... n _ 1
n n n n n2" , ,
2(n-k)k
Hence the required probability = L
n-l
2
k=l n
=-;(n~k- ~k2) =-;{n. (n-1)n _ n(n-1)(2n-1)} = n2 -1 .
n k=l k=l n 2 6 3n
Example 14. An urn contains a white and b black balls, from which k
balls are drawn one by one and they laid aside without noticeing their
colours. Then one more ball is drawn. Find the probability that it is white.
Here the total number of cases of drawing (k + 1) balls from
(a + b) balls is (a + b)(a + b - 1) ... (a + b - k) .
Since the last drawn ball shall be white, we choose one white ball from a
white balls in a ways. Then k balls can be drawn from the rest (a + b - 1)
balls is (a+b-1Xa+b-2)···(a+b-k)ways.
So the total number of favourable cases in which the last drawn ball
is white in (k + 1) drawn is
(a + b - 1)( a + b - 2) ... (a + b - k)a .
Hence the required probability
(a + b - 1)(a + b - 2) ... (a + b - k) . a a
= (a + b)(a + b - 1) ... (a + b - k)
=
a +b .
Example 15. 15 new students are to be evenly distri~uted among 3 classes.
Suppose that there are 3 whiz-kids among the fifteen. What is the probability
that each class gets one whiz-kid and one class gets them all ?
15 students can be evenly distributed among 3 classes in
ENGINEERING MATHEMATICS -llA
20
ibuti . (15)!
di
So the tota 1 numb er 0 f stn utions IS ~.(5!)
(i) We can allot one whiz-kid to each of three classes in 3! ways. Then
the other 12 students can be evenly distributed among 3 classes in
(12)'
l2C x8C x4C ways ==s::i: ways.
4 4 4 (4!)3
So the probability that each class gets one whiz-kid is
3' (12)!
. (4!)3 25
==_.
(15)! 91
(5!)3
(ii) We can allot all the three whiz-kids to one class in 3 ways and the
rest 12 students in
3 x _(~12~)_!
(5!)2 2! 6
==-
So the prob. that one class gets all 3 whiz-kids is (15)! 91'
(5!)3
Example 16. An urn contains 4 white and 6 black balls. Two balls are
successively drawn from the urn without replacement of the first ball. If
the first ball is seen to be white, what is the probability that the 2nd ball
is also white?
Let Al be the event that the first drawn ball is white and A2 be the
event that the second drawn ball is white. Then Al Il A2 be the event
that the both drawn ball is white.
4 3 2
:. P(A1IlA2)==-x-==-
10 9 15
4 2
Also P(A1)==-==-'
10 5
BASIC PROBABILITY THEORY 21
. Let AI' A2 be the event that the ball is drawn from the first and second
.
urn respectively. Then as the urn is chosen by coin-tossing, so we have
1
.
P(AI) =P(A2) =- .
2
Now let A be the event that the drawn two balls are white. Then
5·4
5C2 2.1 5
P (A / Al ) = I2C = 12 ·11 = 33
2 --
2·1
4·3
P(A/ A ) = 4C2 = 2 ·1 = ~
2 6C 6·5 5
2 -
2·1
:. By Baye's Theorem ~A)=~~)~A/~)+~~)~A/~)
1 5 1 2 91
=-.-+_._=-
2 33 2 5 330
Again by Baye's theorem, the required probability
1 5
P(A /A) = P(AdP(A/ Ad = "2' 33 = 25
I P(A) ~ 91'
330
Example 19. A speaks the truth 3 out of 4 times and B 7 times out of 10.
They agree in their statement that from a bag contaning 6 balls of different
colours a white ball has been drawn. Find the probability that the statement
is true.
Let Al andA2 be the events that the joint statement of A and B is
true and false respectively
1 5
Then P(AI) = 6' P(A2) = 6
Now let X be the event that both A and B agreed in their statement.
Then
3 7 21
P(X/A) =-X-=-'
I 4 10 40
BASIC PROBABILITY THEORY 23
(1-x-
P(XjA2)= 1) x(3-x- 1) =--3
4 5 10 5 1000
1 21 5 3 7 1 9
=-x-+-x-- =-+-=-
6 40 6 1000 80 400 100
both are boys, given that (i) the older child is a boy (ii) at least one of the
children is a boy?
Let Al and A2 be the event that the older child is a boy and the
younger child is a boy
1
Then P(AI) = P{A2) = 2
Also Al U A2 = at least one of the children is a boy
and Al nA2 = both children are boys. Since AI' A2 are independent,
• :. P{AI(12)=p{AI)P{A2)=~'~=~
1 1 1 3
P(AI uA2) =P(AI)+P{A2)-P(AI nA2) =2+2-4 =4"'
(i) Thus the probability that both children are boys given that the older
is a boy is
24 ENGINEERING MATHEMATICS -ItA
(ii) The probability that both the children are boys, given that at least
one of them is a boy, is
P(Al uA2)
P(AlnA2) 1/4 1
= P(Al uA2) =3/4 =3'
P(X/X
2
)-
-
--±- - 2..
100 - 25 '
P(X/X)
3
- ~
-
- 2..
100 - 50
L
BASIC PROBABILITY THEORY 2S
Example 22. An urn contains 10 white and 3 black balls, while another urn
3 white and 5 black balls. Two balls are drawn from the first urn and put'
into the second am and then a ball is drawn from the latter. What is the
probability that it is a white ball ?
Let A,B,C be the events that the drawn two balls from the first unr are
both white, both black and one white and one black respectively. Then
P(A) = lOC2 = 10x9 =~
l3C
2
13 x.12 26
P(B) _ 3C2 _ 3x 2 _ ~
- l3C - 13 x 12 - 26
2
lOC 3C 5
P(C) = 1X 1 =
l3C 13
2
When two balls are transferred into the 2nd am, it will contain either 5
white, 5 black balls. or, 3 white, 7 black balls or 4 white and 6 black balls
according to the events A, B, C respectively.
Let W denote the event of drawing a white ball from the second urn.
5 1 3
Then P(W / A) =10 =2'P(W / B) = 10
P(W / C) = -±- =~
1.0 5·
.So the required probability,
P(W) =P(A).P(W/ A)+P(B)P(W/ B) +P(C)·P(W / C)
15 1 1 3 5 2 75 + 3 + 40 59
=-.-+-.-+-.- = =-
26 2 26 10 13 5 260 130 .
Example 23. A student has to answer a multiple choice question with 5
alternatives. What is the probability that the student know the answer, given
that he answered it correctly.
Let B, and B2 be the events that the student knew right answer and
guesses the right answer respectively:Also letA be event that he getstheright
answer. Again let p be the probability that he knew the corrcet answer.
26 ENGINEERING MATHEMATlCS-11A
1
Also P(A / BI) = 1,P(A / B2) =-
5
.. By Baye's theorem,
5p
=
4p+1·
EXERCISES
= P(BC)]
10. If A,B are two events with P(A) = O.4,P(B) = 0.3 and
p(AnB) = 0.2, find P(N nB). [WB.U.Tech 2006 M 302]
[Hints: P(A u B) = 0.4 + 0.3 - 0.2 = 0.5
.: P(Ac (I B) = P(A (I B) -P(A) = 0.5-0.4 = 0.1]
11. If two perfect coins are tossed simultaneously find the probability
of getting at least one head.
12. Out of 120 tickets numbered consecutively from 1 to 120, one
ticket is drawn at random. Find the probability of getting a multiple of 5.
13. If A, B, C are equally likely, mutually exclusive and exhaustive then
find P(A).
14. Find the probability of getting at least one 'Five' from 3 throws of
a perfect die.
15. Two events A and B are such that P(A) = 0.4, P(A uB) = 0.7 and
P(B) = x . For what values of x are A and B (i) mutually exclusive (ii)
independent ?
16.Find the probability that there would be two tail if four unbiased
coins are tossed.
17.Three balls are drawn at random from a box containing 6 red and
4 black balls. What is the probability that the two balls are red and one is
black?
IS.Box A contains 3 black balls and 3 red, box B contains 6 black
balls and 4 red. If a ball is randomly selected from each box, find the
probability that the balls will be of same colour ?
19.5 cards are drawn from a pack of 52 well-shuffled cards. Find the
probability that 4 are aces and 1 is a king.
20.Five persons X; Y,Z, W, S speak at a seminar-lecture. What is the
probability that X speaks immediately before Y ?
21.The nine digits 1, 2, 3 , 4, 5, 6, 7, 8, 9 are arranged in random
order to form a nine-digit number. Find the probability that 1, 2, and 3
appear as neighbours in the order mentioned.
22.Four dices are rolled. What is the chance of obtaining a sum of 18?
23. What is the probability of an odd sum when two dice are thrown ?
24. There are four persons in a room. Find the probability that (i) all
of them have different birthdays (ii) at least 2 of them have the same
birthday (iii) exactly 2 of them have the same birthday. (1 year = 365
days)
25.There are 20 people. Find the probability that among the 12 months
in the year there are 4 months containing exactly 2 birthdays and 4 containing
exactly 3 birthdays?
26. If F(Arill)=~,F(B)=~ and ~A)=..! find ~AuB),P(Ac nff),
352
P(Ac u BC) and P(A nBc).
27.What is the chance that a non-leap year should have 53 sundays?
[Hint. Same as Ex-I]
28. X and Y stand in a line at random with 10 other people. What is
the prob. that there are 3 people between X and Y.
29. Prove that the probability of obtaining six at least once in 4 throws
1
with a die is slightly greater than "2 .
9
[Hints: The required Probability = 1- ( %
2
1
0.52 > ..!. :::
30. What is the probability that a bridge han~ will contain at least one
ace? 48C
[Hints: 1- 52C13 ~ 0.696 ]
13
31.Prove that for any two events A, B
(i) P(A + B) = 1- P(B) + P(AB)
(vi) P(A/B)~P(A)/P(B)
(vii) p( A / B) ~ 1 - P(A / B)
(viii) P(A)/ P(B) = P(A / B)j P(B / A).
30 ENGINEERING MATHEMATICS - UA
2 ( -) 1
34. Given P(A u B) = 3' PAn B = 3"' fmd P(B) and P(A) if
1
P(AIB)="6. .
_ p(AB C) P(ABC)
C) p(AB I C) + P(AB I C) = +--'--:---:-'-
11 P(C) P(C)
p(ABC)+P(ABC) P(AC)
= P(C) = P(C) =P(AIC)].
A SWERS
1. 4"
1 1
6."6 8.
22
5'3'
3
No 11.4" 12.5
1
1
13. 3" 14. 1-
(5)3
"6
15. (i) 0.3, (ii) 0.5
3
16."8 17.
1
"2 18. "21 19. 4/52 C 5
1 1 1
20. 5 21. 72 22. 5/81 23."2
BASIC PROBABILITY THEORY 31
24. (i) 0.984, (ii) 0.016, (iii) 6 x 364 x 363 25. 1.0604 x 10-3
(365)3
23 7 2 1 1
26. 30' 30' 3"' (3 27. -,:; 28. 4/33 30. 0·696
7
32. l3/15, 1/3, 1/5 34. 1/3, 18
colour are left. Prove that the probability that the balls left are white is
a
a+b '
32 ENGINEERING MATHEMATICS-IIA
4. A bag contains 5 white and 4 black balls. If 3 balls are drawn at random,
what are the probabilities of the following:
(i) 2 of them are white
(ii) at most one of them is white
(iii) at lest two are white. [WE. UTech 2007]
[Hints: (i) n(A) =5 C2 X4 C1 (ii) n(A) =5 C1 x4 C2 +4 C3
(iii) 5C2 x4 C1 +5 C3]
5. The probability that a contractor will get a plumbing contract is 23,
and the probability that he will not get an electri contract is 59. If the probability
of getting at least one contract is 4/5, what is the probability that he will get
bothe the contracts?
.6. From the numbers 1, 2, 3, 4, 999, 1000 one number is drawn
-at random Find the probability that the number is multiple of (i) 12 and 18
(ii) 12 or 18
7. A box contains 5 defective and 10 non-defective lamps. 8 lamps
are drawn at random in succession without replacement. What is the
probability that the 8th lamp is the 5th defective?
[ WE. UTech 2005,2007,2008]
5C xlOC X 7'
[Hints: 4 3 •
15 x 14 x 13 x 12 x 11 x 10 x 9 x 8
8. Two sets of candidates are competing for the position of the Board
of Directors of a company. The probabilities that the first and second sets
will win are 0.6 and 0.4 respectively. If the first set wins, the probability
of introducing a new product is 0.8, and the corresponding probability if
the second set wins is 0.3. What is the probability that the new product
will be introduced?
9. An urn contains 2 white and 2 black balls. Balls are drawn
successively at random without replacement. What is the probability that
a black ball appears (i) for the first time in the 3rd drawing (ii) for the
2nd time in the 4th drawing?
10. A packet of 10 electronic components is known to include 3 defectives.
If 4 components are randomly chosen and tested, what is the probability of
finding among them not rrore than one defective?
BASIC PROBABILITY THEORY 33
11. A bag contains 8 red balls and 5 white balls. Two successive draws
of 3 balls are made without replacement. Find the probability that the first
drawing will give 3 white balls and the second 3 red balls.
12. In a bridge game, North and South have 9 spades between them.
Find the probability that either East or West has no spades. (There are only
13 spades in a pack of 52 cards and each player has 13 cards. The players
are designated by the positions they occupy, viz. North, South, East, West.)
13. In a hand of bridge, what is the probability that you have 5 spades
and your partner has the remaining 8.
14. A box contains 7 white and 5 black balls. If 3 balls are drawn
simultaneously at random, what is the probability that they are not all of
the same colour ? Calculate the probability of the same event for the case
where the balls are drawn in succession with replacement between
drawings.
15. An urn contains 3 white and 5 black balls. One ball is drawn and
its colour is unnoted, kept aside and then another ball is drawn. What is
the prob. that it (i) black (ii) white? '
[Hints: Let ff\, Bl be the event that 1st drawn ball is white and
black. Also let fJ1, B2 the event that 2nd drawn ball is white ana black.
3 5 2 3
P(Wd = 8' P(B 1) = 8' P(W2/Wd = 7 ,P(W2/B1) ='7
, 3
.. P(W2) = P(W1)·P(W2/W1)+ P(B1)·P(W2/W1) =-
8
5
:. P(B2) = 8']
16. Each of the two cabinets has, 3 drawers. Cabinet I contains a gold
coin in each drawer, and cabinet II contains a gold coin in one of its
drawers and a silver coin in the other. A cabinet is randomly selected, one
of its drawers is opened, and a gold coin is found. Find the probability
that there is a gold coin in the other drawer.
17. There are three identical boxes, each provides with two drawers.
In the first, each drawer contains a gold coin, in the third, each drawer contains
a silver coin and in the second, one drawer contains a gold and the other a
silver coin. A box is selected at random and one of the drawers is opened If
a gold coin is found, what is the prob. that the box chosen is the second one?
, EM-2A-3
34 ENGINEERING MATHEMATICS-IIA
11111 1
:. P(A)=1· + · + ·O=2 :. P(A2IA)=3]
3 23 3
18. There are three identical urns containing white and black balls. The
first urn contains 3 white, 4 black balls, the 2nd urn contains 4 white 5 black
balls and the 3rd urn contains 2 white and 3 black balls. An urn is chosen at
random and a ball is drawn from it. If the drawn ball is white, what is the
prob. that the 2nd urn is chosen. [w'B. U.Tech 2007]
19. Urn A contains 2 white, 4 red balls; Urn B contains 1 white and 1
red bal. A ball is randomly chosen from urn A and put into urn B .and a
ball is then randomly selected from Urn B. Find (i) the probability that
the ball selected from Urn B is white (ii) the conditional probability that
the transferred ball was white, given that a white ball is selected from
Urn R
20. A box contains 3 types of disposable lights. The probability that
type I light gives over 100 hours of use is 0.7, with the corresponding
probabilities for type II and type ill lights being 0.4 and 0.3, respectively.
Let 20% of the lights in the box are type I, 30% are type II, and 50% are
type III.
(i) Find the probability that a randomly chosen light gives more than
100 hours of use ?
(ii)Given the light lasted over 100 hours, what is the conditional
probability that it was a type II light?
21. Two cards are drawn successively from the pack without replacing
the first. Find the prob. that the 2nd card is also spade, if the first card is
spade.
[Hints. Let A, B be the event that the 1st and 2nd drawn card is
spade .
. P(AB) - 13 x 12 P(A) =~ P(BI A)= P(AB) = 12 =~
.. - 52 x 51' 52 P(A) 51 17·]
22. III a certain class 15% of the students failed in Mathematics, 15%
failed in Chemistry, 10% failed in both Mathematics and Chemistry. A
student is selected at random
L
.J
BASIC PROBABILITY THEORY 3S
10c x 2S 10C 9C 26 -
[ Hint (i) S ,(ii) 1 x 6 x ;][ W.B. U. Tech 2005]
20CS 20CS'
ANSWERS
27 5 10 17 25
1. (i) 0.56 (ii) 28 2. 18 4. (i) 21 (ii) 42 (iii) 42
5 1 1
5. 14/45 6. 0.027, 0.111 7. 429 8.0.6 9. 6'2
10.2/3 11.7/429 12. 11/115 13.26084xlO~ 14.35/44,35/48
5 3 3 1 140 4 1
15. (i) '8' (ii) '8' 16. '4 17. '3 18. 401 19. (i) '9 (ii) 2
2 2 1
20. (i) 0.41 (ii) 12/41 21. 4/17 22. (i) '5' (ii) '3' (iii) 10'
3 20 40 1
(iv) 10' 23. 21; 4i 24. 2 25. 4/25 26.. 09145; .4268
1. The probability that the sum 8 appear in a single toss of pair of a fair
dice is
31 5
(a) 36 (b) 36
5 1
(c) -
6 (d) 6'
2. Of 6 girls in a.class, 3 have blue eyes. If 2 of the girls are chosen at
random, the probability that both have blue eyes is
1 4
(a) - (b) -
5 5
1
(c) 10 (d) none.
38 ENGINEERING MATHEMATICS-11A
3 4
(a) - (b) -
4 3
1 1
(c) - (d) -
4 3
5 7
(a) 12 (b) 12
1
(c) 6" (d) none.
(a) P(A))+P(A2)-P(A).nA2)
(b) P(A1)·P(A2)
(c) P(Ad+P(A2)
(d) P(Ad - P(A2).
1
(c) - (d) none.
2-
11. A coin is tossed successively three times. The probability of getting
exactly one head is
1 1
(a) - (b) -
8 2
3
(c) - (d) none.
8
12. If P(A) = 0.2,P(B) = O.4,P(A u B) = 0.6, then the events A,B
are
22. If P(A) = ~,P(B) == ~,P(AB) = : ' then the value of P(A uB)
IS [WB.U.Tech 2007]
6 3
(a) - (b) -
7 7
7
(c) 1 (d) 12'
26. 3 balls are drawn from a bag containing 6 red and 5 white balls.
The probability of getting the balls all red is
6 3
(b) -
(a) 11 7
1 1
(c) 36 (d) "3'
28. A bag contains 5 brown and 4 white balls. A man pulls out two balls
from the bag. The probability that they are of the same colour is
1 3
(a) - (b) -
5 7
4 (d) none of these.
(c) -
9
29. Three of six vertices of a regular hexagon are chosen at random.
The probability that the triangle with three vertices is equilateral is
1
1 (b) -
(a) - 3
6
1
1
(c) - (d) 10'
7
30. An unbiased coin is tossed repeatedly. If 'tail' appears on first four
tosses, then the probability of 'head' appearing on the fifth toss is
1
1 (b) -
(a) - 4
2
2
1
(c) - (d) g'
3
BASIC PROBABILITY THEORY 43
31. Seven white balls and three black balls are randomly placed in a
row. The probability that no two black balls are placed side by side is
3 7
(a) ~ (b) 15
2 1
(c) -
9 (d) "4.
32. Five boys and three girls are seated at random in a row. The probability
that no boy sits between two girls is
3 1
(a) 28 (b) 28
3 4
(c) 25 (d) 19.
33. Three different numbers are selected at random from the set
{1,2,3,4,5,6, 7,8,9,1O}. The probability that the product of two of the
34. The probability of a student getting first class, second class and
121
third class at an examination are 10' 5 and 5· respectively. The probability
that he fails is .
1 4
(a) 10 (b) -
9
3 7
(c) 10 (d) U.
35. The probability that Pradyut passes is 0.9 and Subrata passes is
0.8. The probability that at least one of them passes is
-
ENGINEERING MATHEMATICS -llA
44
36. A fair die is thrown. The probability that either an odd number or a
number greater then 4 will turn up is
3
2 (b) -
(a) - 7
5
2
2
(c) - (d) 3' .
7
37. Three bombs are dropped to destroy a bridge. The probabilities of
hitting the bridge by these bomb are 0.5, 0.2 and 0.1 respectively. The
P(BnC) =
1
3
(a)- (b) 14
5
1 3
(d) ;;.
(c) 12
40. The probability that at least one of the events A and B occurs is 0.6.
If A and B occur simultaneously with probability 0.2, then p( A) p( B)
+ =
41. Two numbers are chosen from the set {1,2,3,4,5,6} one after
another without replacement. The probability that the smaller value of the
two is less than 4 is
4 1
(a) - (b) -
5 5
3 7
{c) -
5 (d) 11.
ANSWERS
:.P(A;)= nciiqn-i
Note. Probability of all's' in En is P(An) = p" and probability of all 'f
in En is P{Ao) = qn .
Remark. In a subsequent chapter we shall see this Biriomial Law has
an important role in the developlment of a special type of discrete
distribution called Binomial Distribution. In this book, in fact most of the
problems which could be discussed under this law will be presented in
that section of 'Binomial Distribution'.
Example 1. Let E be the experiment of drawing a ball from an urn
containing 5 white, 7 black, 3 blue and 8 red balls. In a draw let s = 'the
ball is white or blue'.
8 8 15
:. p =P(s) =- So q=I--=-
23 . 23 23
If this experiment E is repeated 9 times,
Probability of four success( among these 9 trials)
........................... and so on .
Since pes) =p, P(f) =q and peS) =1
so, PC AI) = p x 1x 1x 1· =p
P( A3) = q . q. q .q .p ·1.1 = P q4
P( A4) = q .q .q .q . q . q . p .1 = P q6
.. ...... and so on .
:. Probability that A wins
= P(AI U A2 U A3 U········..)
= P( AI) + PC A2) + PC A3) + .
2 4 6
=p+pq +pq +pq + .
= p(l+q2 +q" +q6 + )
."
1
= P . 1- q2 .: q2 common ratio of the G. P
• 1 36 6
=-.---=-.-=-
6 1-(%Y 6 11 11
EM-2A-4
50 ENGINEERING MATHEMATICS - rIA
Solution. Let El' E2' E3 be the experiments that the problem is given to
. .
A, Band C respectively.
1 1 . 1
Now, peA) = 2' PCB) = 3" and P(C) = 4"
- 1 1 . - 1 2 - 3 .,
:.P(A)=l--=-, P(B)=l--=-and P(C)=-
2 2 3 3 4
We consider the joint independent experiment (EI' E2' E3) .
=(A,B,C)
1)3(6)17 617
or, the required probability = 2OC3
( -7 -7 = 1140 x -20
7
Example 3. In five throws with a coin find the probability of
(i) 3 heads (ii) at least 3 heads (iii) at most 3 heads
Solution.
Let E be the experiment of throwing a coin. 's' = Head 'I" = Tail
111
:,p="2,q=I-"2="2'
We consider the finite Bernoulli trial of 5, E5 .
(i) By Binomial law probability of3 heads
5
=P(A3)=5C3P3q5-3=5C3(kJ(kY =1 6
= p( A3 U A4 U As )
= 1- p( Ao U Al U A2) = 1- {5 CopOq5-0 + 5Clpll-1 + 5C2p2q5-2}
=1-C I
2 + :2 + ~~)=k
(iii) Probability of at most 3 heads
= P(AoU Al U A2 U A3) = P(Ao U Al U A2) + P(A3)
1 5
=- +- using (i) and (ii)
2 16
13
=
16
Example 4. A can hit a target 4 times in 5 shots; B 3 times in 4 shots; C
twice in 3 shots. They fire a target. What is the probability that at least
two shots hit ?
Solution : If EI' E2 and E3 be the experiments that A, Band C fire the
target respectively then we can treat this as a joint independent experiments
( EI , E2' E3) •
Let AI> A2' A3 be the event that A, B, C hit the target respectively.
52 ENGINEERING MATHEMATICS -IIA
Then
- 4 1
:.P(Al)=I--=-
5 5
- 3 1
:.P(A2)=I--=-
4 4
- 2 1
:.P(A3)=I--=-
3 3
Now the event 'at least two shots hit'
. ::::;(41, A2,A3) U (AI ,A2,A3) U (AI ,A2,A3) U (AI,A2,A3)
'.. :;
x = {(3,5),(5,3),(2,6),(6,2),(4,4)} :. n(X) =5
Y = {(2,3),(3,2),(4,1),(1,4)} :. n(Y) =4
5 4 1
. P(X) -
.. - 36'
P(Y) - -
- 36-"9
.,' n(S) = 6 x 6 = 36
THE BERNOULLI TRIAL 53
:. P(X) =1-2. =~
36 36
:. P(Y) =1-i =%
We think this is an infinite sequence of Bernoulli trial Eoo' Here the
event 'A wins'
= {X ,S,S,.... } U {X ,Y ,X ,S,S,.... } U {X ,Y ,X ,Y ,X ,S,S,.... } U· .....
(S is certain event)
So, the probability of A wins in the game
.
= P(X) + P(}()P( Y)peX) + [Since the successive trials are
Bernoulli independent trial]
l
5
=3 6 + ~~.%.:6 +G~J (%J. :6 + .
, 5
= 36 45
8 31 =76 [ .: Thi's is an infini
Illite GP senes
. wiith c.r"9·36<1
8 31
1--·-
9 36
?
Example 6. A missile hits target with probability 0.3. How many missile
should be fired so that there is at least an 80% probability of hitting a
target?
Solution. Let E be the experiment of firing a missile.
Let E be repeated independently n times
Let's' = 'a missile hits the target'
:.p=P(s)=O.3. So q=P(f)=1.1.0.3=0.7
In the Bernoullis n trials, probability of failing to hit the target in each
trial = qn .
. " .
Probability of hitting at least once
= 1- Probability of failing each trial = 1- qn = 1- (0.7)" .
We have to find n such that P(An) > 0.8
=(~J +2'~'~= ;5
Example 8. A factory produces blades among which 20% are defective.
If 5 blades are drawn at random from a day's production, find the
probability that there will be (i) exactly 2 defectives (ii) not less than 2
defectives.
Solution. Let E be the experiment of drawing a blade from the factory's
huge production.
THE BERNOULLI TRIAL 55.
and P(AI)=
5
Cip q
I 5-1
=5x x
S
·1 (4)4
S = (4)4
S
Required probability = 1- P(Ao U AI)
= l-{P(Ao) + P(A])}
321
Then, peG,) = 4' P(G2) = 4' P(G3) = 4
• 123
PCB,) =-, P(B2) = -,P(B3) =-
444
We consider the joint independent experiment (E" E2, E3)
The event 'I girl 2 boys'
= {(G ,B2,B3
1 ),(B, ,G2,B3 ),(B, ,B2 ,GJ)}
:. the required probabili.ty
= p( G1 ,B2,BJ) + P(B1,G2,B3) + P{B"B2,G3)
= P(G,)P(B2)P(B3) + P(B,)P(G2)P(B3) + P(B,)P(B2)P(G3)
3 2 3' 1 2 3 1 2 1
=_._._+_._._+_.-.-=- 13
4 4 4 4 4 4 4;' 4 32 .
Example 10. In an infinite sequence of Bernoulli trial with probability of
1
success '3' find the probability that 7 failures will precede the first success.
1 1 2
P(s) = '3,P(f) =1-'3 ='3 and peS) "" 1.
=~.~.~.~.~.~.~.!.1.1.1 .
3 3 3 3 3 333
_(2)7 1_ 27
- '3 ''3-3"8
Example 11. Three persons A, Band C toss a coin in succession and the .
first to obtain a head wins the game. Find the probability of winning of
A, Band C respectively.
Solution. E be the experiment of tossing a coin. In a single trial
's' = 'Head'·
:. p = pes) =! in a trial. Then q = P(f) =!
2 2
L
THE BERNOULLI TRIAL 57
=(p ·1·1··· ..)+(q. q .q. p·I·I·· ...) +(q. q- q.q. q. q- p·I·I··· ..)+ .....
= pCI + q3 + l + q9 + (0)
= qp(1+ l + q6 + 00)
1J
111 1 11182
=qp--=-.-. =4"·--1 =-.-=-
1- q3 2 2 1_ ( 1- "8 4 7 7
2 36 21111
=q p(l+q +q +·········uptooo)=q p--=_._._-
I_q3 22 2 1-!
1 8 1 8
=-0-=-
8 7 7
Example 12. A player repeatedly throws a coin and scores one point for
a Head and two points for a Tail. He stops throwing whenever he scores
a total of 4 point. Find the probability of scoring 4.
Solution. In fact this is a finite Bernoulli trial with success,
H = Head and failure, T = Tail,
Now P(H) = 1/2, P(T) = 1/2 , in a single trial.
U{T,H,H} U{T,T}
. d pro b abili
:. th e require 1 1 1
ility =-+-+-+-+-=- 1 1 11
16 8 8 8 4 16
THE BERNOULLI TRIAL 59
U{Ao,Bo,Co,Ao,Bo,Co,Ao,S,S, } U· upto 00
= {P(Ao)P(S)P(S) } + {P(Ao)P(Bo)P(Co)P(Ao)P(S) }
= -211321231232
+ -. -. - .- + _. -. -. - .- .- .- + .....
.
upto 00
332433343343
=~+t·~+(tr·~+(tJ.~+..... upto eo
,.
L
60 ENGINEERING MATHEMATICS-IIA
Exercise 2
1. Three cards are successively drawn from a full pack, the card
drawn being replaced every time. Find the probability that first card is
spade, second card is heart or diamond and the third card is queen.
2. Four cards are drawn successively from a pack of 52 cards with
replacement. Find the probability that all the four cards are of the same
suit.
.
[Hmt. (13 13 13 13) x4]
_._._0.-
52 52 52 52
3. The probability that Ashok can solve a problem in Business Statistics
423
is 5"' that Arnal can solve it is "3' and that Abdul can be solve it is "7 . If
all of them try independently, find the probability that the problem is solved.
4. A candidate is selected for 3 posts. For the first post there are 3
candidates, for the second post there are 4 and for the third there are 2.
What is the chance of getting at least one post ?
[This is complement of 'getting none of three'.
8. There are three men aged 60, 65 and 70 years. The probability to
live 5 years more is 0.8 for a 60 year old, 0.6 for a 65 year old, and 0.3
for a 70 year old person. Find the probability that at least two of the three
persons will remain alive 5 years hence.
[Hint. (0.8)(0.6)(1- 0.3) + (0.8)(1- 0.6)(0.3)
+(1- 0.8)(0.6)(0.3) + (0.8)(0.6)(0.3) ]
9. A and B toss a coin alternately and the first to obtain a head wins
the toss. If A starts the game, find the probability of his winning
.
[Hint. the required probability = L:"2
(1 )2n ."21 "3]2
=
10. A player repeatedly throws a coin and scores one point for a Head
and two points for a Tail. He stops throwing whenever he scores a total
of 5 point. Find the probability of scoring 5.
U. In an infinite sequence of Bernoulli trial with probability of success
1
4" ' find the probability that 6 failures will precede the first success.
12. Two persons A and B throw a die alternatively and A starts
throwing till one of them gets a multiple of 3 and wins the game: Find
their respective probabilities of winning.
[Hint -1 + (2)2
- .-1 + (2)4
- .-1 1- ..... ·00 ]
·3 3 3 3 3
13. A man alternatively tosses a coin and throws a die begining with
the coin. Find the probability that he will get a head before he gets '5 or
6' on the die ?
J
1 1 2 1 1 .2 1 2 1
[Hint - + - x - x - + - x - x - x - x - + ...... 00 ]
·223223232
>
Answers
1 1 101 3 80
1. 104 2. 64 3. 105 4.4" 5. 243
2 21
6. ·624; 4 7. 0·96 8. ·612 9. "3 10. 32
36 2 3
11. -7
4
12. "5 13.4"
r
~3
~
I~ _I_S_C_RE
D __T_E_RA ND__O_M_V__MU
__A_B_L
ANn ITS EXPECTATION
__
E
The range of the function X i.e. the set of all values assumed by X is
called the Spectrum of the random variable.
Discrete Random Variable. A random variable (r.v) X is said to be
discrete if the spectrum of X is fmite or countably infmite i.e an infinite
sequence of distinct values.
F(x)=O, x<O
1
= O~x<l
4'
1 1
=-+- 1~x<2
4 2'
111
=4+2+4' z s e.
Properties of Distribution Function.
(iii) F(x) is continuous on the right at all points and has a jump
discontinuity on the left at x = a, the height of jump being equal to
P(X = a) i.e., lim F(x) = F(a) and
x-+a+
F(a) - lim F(x) = P(x = a)
x-+a-
(iv) Suppose a and b are any real numbers such that a < b
Then P(a < X ~ b) = F(b) - F(a) ,
P( a < X < b) = F( b) - F( a) - P(X = b)
and P(a ~ X < b) = F(b)-F(a)-P(X = b) + P(X = a)
Proof: Left as exercise.
Illustration. (i) Let X be a random variable denoting the number of points
appearing in a toss of a die. The distribution of X is
X 123 4 5 6
1 1 1 1 1 1
6 6 6 6 6 6
Now, if x c l . F(x)=P(X~x)=O
If 1 s x < 2, F( x) = P( X s x) = fl = ~
112 --
If 2 s x < 3 , F( x) == P(X s x) ==fi + f2 ==6' + 6 == 6'
and so on.
5/6
4/6
3/6
2/6
1/6
x
1 2 3 4 5 6 7
From the graph it is clear that F( x) is a step function, non decreasing
and is continuous on right at x = 1, 2, 3, , 6 and has a jun
1
discontinuity on left at 1,2,3, ,6, the height of jump is "6.
= 3C3/8C3 =~
56
t, = P(X = 1) = Probability of one white ball
X o 1 2 3
1 15 15 5
56 56 28 28
From this we get the probability of the event like 'at most two white
= fo + t. + 12 = 1- f3 = 1- ~28 = 23
28'
3.4. Expectation or Mean of a Discrete Random Variable
Let X be a discrete random variable whose distribution is
X
t,
Then the mean or expectation or expected value of X denoted by
E(X) or m(X) or simply m is defmed as
So the distribution of X is
X 2 3 4 5 6
1 1 1 1 1 1
t. .- - - - - - -
6 6 6 6 6 6
68 ENGINEERING MATHEMATICS -IIA
Properties of Expectation
(i) E( a) = a, where a is a constant.
Illustration. A number is chosen at random from the set {I, 2,· ..,l00}.
We are to find the expected value of the chosen number X and the
expection of 3X3 + 2.
Let X = first number and y = second number. So the distribution
. of Xis
X 1 2 3 100
1 1 1 1
100 . 100 100 100
So,
1 1 1 1
E(X) = -·1+-·2+···+-·100 = -(1+ 2+3+···+100)
100 100 100 100
1 100(100+1) 101
=_. =-
100 2 2
DISCRETE RANDOM VARIABLE & ITS EXPECTATION 69
Thus o = +~var(X) .
Remarks: (i) The variance describes how widely the probability masses
are spread about the mean i.e it gives an inverse measure of concentration
of the probability masses about the mean which is called the measure of
dispersion.
(ii) As Var(X) = 0 only when X - m = 0 i.e, X = m, so in that
case the whole mass is concentrated at the mean.
Theorem.
2
(i) Var(X)
E(X2) _ m2 = E(X2) _ {E(X)}
=
= E{X(X -1)}-2m.m+m+m2
=E{X(X-l)}-m(m-l.
Note: In fact the result (i) and (i~ of the
...-/ bove theorem are used to
evaluate variance and standard deviation.
3.6.l\'Ioments of a Random Variable
The r-th moment of a r.v X about A denoted by J.l; is defined as
J.l; = E(Xr)
Ther-th moment about the mean of X is
u, = E (X - xy) which is known as r-th Central Mo~ent of
X, where X is mean of X.
Example. Let Xbe a discrete random variable whose distribution is
X: 3 6 7 11
J.l3 =E(x-5.7i)
= (3 - 5 . 7)3 x 0 . 5 + (6 - 5 . 7)3 x 0 ·1 + (7 - 5 . 7)3 x 0 . 2
+ (11 - 5 . 7)3 x 0 . 2
Theorems:
(1) o-th moment about any number is 1
(2) 1st moment about 0 is mean of the variable.
(3) 1st central moment is always zero.
(4) 2nd central moment is variance.
Proof. Let X be the random variable.
(1) P2 = p~ - p:2
=E(X2)_X2 =E(X2)-{E(X)}2
x o 2
1 1
2 23
Example. 2. A random variable X has the following probability mass
function [ WB. U.Tech 2007]
X 01234567
P(X=k)=f(x) :
(i) Determine the constant k
(ii) Evaluate ~X <6), ~X2!6),~3<X ~6) and ~3<X/X~6)
=1_{2(~)2
10
+7.(~)2+~}
10 10
=~
100
81 19
.. P(X ~ 6) = 1 ~ P(X < 6) = 1- 100 = 100
33
P(3<X~6)=P(X=4)+P(X=5)+P(X=6) = 100
1 1~x<2
10
2..+2.2..=~
10 10 10
3 1 1
-+2·-=- 3~x<4
10 10 2
1 1 4
-+3·-=- 4~x<5
2 10 5
4 (1)2 81
5+ 10 =100
5~x<6
81 + 2.
100
(2..)2
10
= 83
100
6sx < 7
-oo<x<O
Example. 3. Let F(x) =0,
1
=- o s x <1
5'
3
= 1~x<3
5'
=1, 3~x<ex:>
Show that F(x) is a possible distribution function. Detennine the spectrum
and the probability mass of the distribution. [w.B.U.Tech 20031
Hence find the mean and standard deviation of X.
Clearly F(x) is monotonic non-decreasing function and is continuom
:.Var(X)=E(X2)-{E(X)}2=4- (8)2
-
5
36
=-
25
given that P(X = 0) =.! and P(X > 1) = ~. [W.B. U.Tech 2004]
6 3
ENGINEERING MATHEMATICS - UA
76
We have F(-oo) =0
:. lira F(x) =0
i.e., lim (A) =0 .. A =0
x-+-oo
x-+-oo
i.e.
,
lim (D)
X~OO
=1 :. D =1
[':P(X=a)=F(a)-lim x-+a-
F(x)]
:. C - B
. =..!:..6 .., (1)
Again P(-oo < X < 00) = P(-oo < X ~ 1)+ P(l < X < 00)
2
:. 1 = P( -00 < X ~ 1)+ P(X > 1) = P{ -00 < X ~ 1)+ "3
2
:. P{-00 < X ~ 1) = 1- -
3
i.e., "F(l)=..!:.. [.: F(x)=P(-oo<X~x)]
3
.. C=..!:..
3
1 1 1 1
.. From (1) B=C--=---=-
6 3 6 6
:. A = 0, B =..!:.., C =..!:.., D =1
6 3
are marked by the number O,..!:..,~ , ..., n -1 ,!!:. . If X denotes the number
n n n n
shown then fmd (i) the expectation of X (ij) standard deviation of X
1
Note that probability of each face is --1 0
n+
So, the distribution of X is
1 2 n-1 n
X 0
n n n n
1 1 1 1 1
t.
n+1 n+1 n+1 n+1 n+1
t
Obviously X assumes values Xi = -n 0
00 ax = ~~;:
Now,
1 (3 3 3) 1 {n(n+1)}2 n+1
= n3(n+1) 1 +2 +ooo+n = n3(n+1) 2 =-4-n-
Hence E( X - ~ r =0 0
78 ENGINEERING MATHEMATICS -11A
-t" '(1
00
=e L...J~ -e -t
'1
)t- -t"·
co
=e L...J~Zi-I
i=O i=O
1 1 1 1 1 .1 21 7
Now, E(x) = 1'6+2'6+3'6+4'6+5'6+6'6 =6= 2"
Expected amount of money = E( 2x + 5) = E( 2x) + E( 5)
7
= 2E( x) + 5 = 2 x 2" +5 = 12,
Example 9. Find the second, third and fourth central moments of X where
P(X = I) = a, P(X = 0) = I-a,
Solution. Obviously the distribution of X is
X : 0 I
/; : I-a a
Now, (X = 0) = {T,S,S, =l
:. P(X = 0) = peT) peS) P(S)··········
=.!..1.1 =.!.
2 2
(X = 1) = {H,T,S,S, oo}
:. P(X = 1) = P(H)P(T) P(S) P(S) .
=~'~'I..I..·..·=(~r
Similarly P(X = = 2) P(H) P(H) P(T)P(S) =(~J
Thus P(X=3)=(~r 'P(X=4)=(~J ,. -and so on
= p 2 + 2 p 3 + 3p 4 + .
(putting p = -)1
2
DISCRETE RANDOM VARIABLE & ITS EXPECTATION 81
2
=p (l-p)
-2
=2"2.
1
( 1-2 1)-2 ="4·2
1 ()-2
1 .1
="4x2
2
=1
EM-2A-6
-----
ENGINEERING MATHEMATICS -llA
82
_ 11121 n 1
X=E(X)=O-+_·_+_·_····+_·-
n+1 n n+1 n n+1 n n+l
n i 1 1 n
=2,-'-=
n n+1i=O n(n + 1) i=O
2,i
1
= (1+2+3+ ... +n) =
1 1 n(n+ ) ==,~
n(n + 1) n(n + 1) 2 2
2 3 n(n + 1)(2n+ 1)
1 n(n + 1)
} 2n2(n+l) 6
== n3(n+l)' { 2
3 n(n + 1) 1
+---~-~
4n(n+l) 2 8
m r 1.E(
3. If the random variable X has mean 'm' and S.D o show that
E( X : = X : m) = 0.
2
1. 1.0,2.9
2'n(n+1)
4. X 1 2
t,
°25
--
10
-
1
36 36 36
5. X 1 2
°--
188 32 1
t. -- --
221 221 221
ENGINEERING MATHEMATlCS-llA
84
6. X 0 2 3 4
1 5 10 5 1
Prob. : - - - - -
42 21 21 21 42
7. 2 8. 1.4
[II1 LONG ANSWER QUESTIONS
x :« -2 -I o 1 2 3
P(x) 0.1 k 0.2 2k 0.3 3k
(i) Find k
(ii) Evaluate P(X < 2), P(X s 2), P( -2 < X < 2) [W.B.U.Tech 2006]
(iii) Determine the distribution function F( x) of X
[W.B. U.Tech 2005]
6. The spectrum of the random variable X consists of the points
1,2" ", n and P(X = i) is proportional to .. 1 . Determine the distribution
.
function of X
~(~+1)
kn n+1
or, -- = lor, k = --]
n+1 n
~
P(3<X~n)=L./i=k~ ~(1 -;---. 1)
- n+1(1---
=- 1) n-3
=-
4 4 z ~+1 n 4 n +1 4n
expectation.
11. Two players A and B agree to play a game under the condition
that A will get from B Rs. 3 if he wins and will pay B Rs. 3 if he losses.
The probability of Pt:s winning is p. Find the mean and variance of A's
gain. [Hints:.X= A's gain =3, -3, p(X=3)=p, p(X=-3)=1-p]
12. Evaluate the expectation of the number of sucess preceding the
first failure in an infmite sequence of trails with probability of failure is p,
where each trial gives two possible outcomes - success or failure.
[Hints: X = 0,1,2,'" ... up to co and P(X =i)=(l- py pl·
13. The first two moments of a discreet r.v about 5 are 2 and- 20
respectively. Find the mean and variance of the r.v.
14. The first two moments of a random variable about 4 are
-1.5 and 2.7 . Find the first two moments about zero. Also find the mean
and standard deviation of the r.v.
DISCRETE RANDOM VARIABLE & ITS EXPECTATION 87
15. The first four moments of a discrete random variable about 1 are
2·6, 10·2,43·4 and 192·6 respectively. Find the first four moments of
X about 4.
16. The first, second and third moments of a random variable about
2 are 1, 16 and -40 respectively. Find the mean, variance and the third
central moment.
Answers
1 1 1 8 7
1. (a)(i) k = ~
6
(ii) "2 (b)(i) k--
- 81' (ii) 9'9' 27' (iii) 6
2. X 0 1 2 3
1 3 3 1
fi - - - -
8 8 8 8
F(x)=O, x<O
1
=- O~x<l
8'
1
= 1~x<2
2'
7
= 2~x<3
8'
= 1, x~3
3. (i) X: 0 2 34
E(2X - 9) is
(a) 10 (b) 1 (c) -4 (d) -9.
2
7. When the variance of a random variable is "3' the •
-1 o 2
The statement is
1 1 1
11. If P(X = -1) = 4"'P(X = 0) = 2'P(X = 1) = 4 and
5
(a) ] 0 (b) 4 (c) 6 (d) "6
1 1 1
12. If P(X = -1) = 4"'P(X = 0) = 2'P(X = 1) = 4" and
111
P(Y = -2) = 3'P(Y = 10) = 3'P(Y = 4) = 3 and if X,Y are
independent then E(XY) =
(a) 10 (b) 4 (c) 0 (d) 1
13. Four coins are tossed. Expectation of number of heads is
7 1 2 4
(a) 15 (b) 15 (c) 15 (d) 15
1 2 1
17. For the pmf f(x) defined by f(O) = "7,f(I) = "7,f(2) ="7
2 1 6 .
f(3) = "7,f(4) ="7 the value of P(X ~ K) ="7' Then K =
20. Throwing two unbiased coins simultaneously Diku bets Buku that
he will receive Rs 4 from Buku if he gets 2 heads and he will, give Rs 4 to
Buku, otherwise. If P(x) be the pmf of Diku's gain then P(Diku's Rs 4
loss)= 1
1 4 3
(a) -
4
(b) -
7
(c) -
4
(d) '5'
21. For a random variable X, E{(X _1)2} = 10 and
23. Let X be a discrete random variable assuming values l,2,3,4, ... and
co
1 P
(d) -
(b) p+1 q'
DISCRETE RANDOM VARIABLE & ITS EXPECTATION 93
A SWERS
1.a 2.b 3.d 4.c S.c 6.b 7.c
8.b 9.a 10.b ll.b 12.c 13.b 14.d
. 15.b 16.c 17.a 18.d 19.d 20.c 21.d
4.1. Introduction:
A discrete random variable is said to have a special type of distribution
if its pmf gets a special form.
Two such type of distributions are discussed in this chapter. Definition
of each of the distribution is first given and their properties and field of
fitness are illustrated in this chapter.
4.2. Binomial Distribution.
A discrete random variable X is said to have a binomial distribution
with parameters p( 0 < p < 1) and n (a positive integer) if its distribution
is given by
X o n
10
i = 0,1, 2,.··,n
For instant, one probability mass
+
11
(nlrl r:
= 1r1-P
nl
= np(l- p) - etc.
= ~~!
(:)(~r(~r+(:)(~YG)+(:)(~r =
96 t:NGINEERING MATHEMATICS - IlA
=npt(~-I)pi-l(l_ pt-i
i=l L-l
=npr=0
-l n-
r
I} r
(l-p)
n-l-r
, replacing i-I by r
~
n-l
=np (p + 1- P) = np
(ii) Now
i
E{X(X -I)} = ti(i-l)(~)pi(l- pr-
i=O L
=n(n-l)p2~ n (ni-2
- 2)pl-2(1_
. p) n-i
=n{n-l)p
2 -2
r={)
n-
r
2}r (l-p) n-2; ,replacing i - 2 by r
~
n
= :L{i(i-I)+i} "cipi(l- p),,-i
i=O
" n
= :Li(i -1) "ci/ (1- »r: + :Li "c.p' (1- p),,-;
;=0 /" i=O
_~·(·_I)n(n-l)n-2
- ~I I . . Ci-2P
i(l_
P
),,-i ~."
+ ~l eiP
'n- P ),,-1
i=O 1(1 -I) i=O
EM-2A-7
l ~ ~
98 ENGINEERING MATHEMATICS -IIA
n n
= n(n -l)P2L n-2Ci-2p;-2(1- pt-; + Li nc;/(l- p)n-;
;=0 ;=1
n-2
=n(n-l)p22: ,,-2Cjpj(l- p),,-2-j +E(X) [putting }=i-2]
j=O
= n(n _1)p2 {(l- p) + p}"-2 + np using previous theorems result
=n(n-l)p~ +np
(i ii) The third order moment about zero,
" {iO
=L -I)(i - 2) + 3i(i -I) + i} "c.p' (1- p)"-;
;=0
n
= n(n -l)(n - 2)iL n-3c;_3/-3(1_ p),,-;
;=3
n
+3n(n _1)p22: n-2c;_2/-2(1- p)n-; + E(X)
;=2
,,-3
=n(n-I)(n-2)p3 ,,-3cjpj(l_p),,-3-i
L
j=O
,,-2
+3n(n _1)p2 L n-2Ckpk (1- p),,-2-k + np
k=O
(ii) Using the relation among central moments and moments about
any number we get
The third central moments,
, 3 ' , 2 '3
P3 = P3 - P2 PI + PI
= n(n -l)(n - 2)p3 + 3n(n _1)p2 + np
-3{n(n-l)p2 + np}np+2(np)3
= np(2p2 - 3p + 1) = np(l- p)(l- 2p)
(iii) Left to the readers as exercise.
r
Illustrative Examples
Example 1. The mean and s.d of a binomial distribution are respectively
, 4 and J% .Find the values ofn and p.Hence evaluate P(X = 0).
[ W.B. U. Tech 2006]
We know the mean and s.d of a binomial variate are respectively np
r and ~np(l- p).
. 8
.. np = 4 or, np(l- p) = '3
8 1
.. 4(1- p) = '3 ~ p = '3
.. n = 4x 3 = 12
O( 1-~ )12-0 = ( ~ )12 .
P(X= 0) = fO=12CO ( )
~
_6 (1)0(1)6
- CO"2
6 (1)1(1)5 6 (1)2(1)4 6 (1)3(1)3 _ 21
"2 + C "2 "2 + C "2 "2 + C "2 "2 - 32.
1 2 3
be defective is
1
10.
If 12 such pens are manufactured, find the prob. that
(i) exactly two will be defecive (ii) none will be defective (iii) at least two
will be defective
= 10 .
1
.'. Prob of success in a single trial p
L
SPECIAL TYPE OF DISCRETE DISTRIBUTION 101
12X 1
(ii) required probality P(X = 0) = ( 0 10
)O( 1- 101 )12-{)= (910)12
= 0.2833·
11
(iii) P(X=I)= 12)( - 1 )1( 1-- 1 )12-1 =12x- 9 =3755.
( 1 10 10 1012 .
1 2
and q =1-3'='3
1
:. Mean = E(X) = np = 3 x - = 1
3
T
1 2 2
and variance = npq = 3 x - x- =-
3 3 3
Example 9. Find the probability distribution of the number of boys in a
family with 3 children, assuming equal probabilities for boys and girls.
Graph the distributions. Also fmd the distribution function F(x) for the
random variable X.
:; Let E be the experiment of picking a child in the family.
The event 'boy' = success.
We have p = probability of boy
1
="2 by hypothesis. E is repeated 3 times.
Let X denote the number of boys.
Then X can assume the values 0, I, 2, 3. X has the Binomial
distribution.
X 0 I 2 3
t. to t. f2 f3
where t, = G)(~}(~r-i(~)(~rG)~
= =
1/8
--~--~~--~------6-------~--~x
Using the above table we obtain the distribution function F( x) as
= 1, 3::s;x<oo·
Example 10. Suppose that half the population of a town are conswners
of rice. 100 investigators are appointed to find out its truth. Each
investigator interviews 10 individuals. How many investigator do you expect
to report that three or less of the people interviewed are consumers of
rice?
•
·r
= (1)10
2 +10x21(1)9
2 +4\2
ll)2(1)82 j1)3(1) 2
+121..."2
1)10 (1 )10
= (2 {I + 10 + 45 + 120} = 176 x 2
.. np=2 (1)
4
np(l- p) ="3 (2)
Solving (I), (2), we get
1 1 2
p = 3" ,n = 6, q = 1- p = 1- 3" = 3" .
fo = P(X = 0) = Co
6 (1)0(2)6
3" 3" = 729
64
t. = P(X (1)1(2)5
= 1) = 6 C1 -
3
-
3
= - 64
249
._-
Thus the required distribution of X is
X : 0 I 2 3 4 5 6
64 64 80 160 20 4 1
. 729 243 243 729 243 729 729·
1
Example 12. The probability of a man hitting a target is "3
(a) Ifhe fires 5 times, what is the probability of his hitting the target at
least twice?
(b) How many times must he fire so that the probability of his hitting
the target at least once is more than 90%.
= 1-P(X = O)-P(X = 1)
= 1_ CO(%r(~r Cl(%}(~r
5
_5
131
= 243 .
(b) Let n be the smallest number of fires so that the probability of hitting
the target of least once is more than 90%.
By the condition
fo t,
e-~Ili
where the prnf, t. = P( X = i) = -. t.
I-
e -~Il i
Note : (1) Since the parameter 11 > 0, t. = -,-,t. - ~ O.
108 ENGINEERING MATHEMATICS -IIA
=e-~t; ao
(i-1)!
Ili
=e-~ (Il
3
112 11 114
0!+T!+2T+3f+···uptooo
J
= II II £
,..e-~ (1 + ,.. + 2! + ~3! + ... up to ooJ = Ile -~ . e ~ = Il
t;i(i -l)e-~ ~!
i ao
= e-~1l2t; (r-
i-2
2)!
= e-~1l2(1+.e.+£+~+ up to ooJ
1! 2! 3!
ENGINEERING MATHEMATICS -UA
110
So, by an earlier result, (See the Theorem of Art 2.6) the variance,
Var(X) == E{X(X -I)} - m(m -1) ,where m == mean
== 112 - 1l(1l - 1) == Il
Note. In light of the above theorem we considered the parameter Il == At
in our previous illustration.
Moments of Poisson Variate
In the following two theorems we find the moments about zero and
the central moments of a random variable X having poisson distribution.
Theorem 1. If X has poisson distribution with parameter J.l then its
co -Jl i
== L{i(i-l)+i}~
;=0 II
<Xl e-Jl J1.; ~ e-Jl J1.; co e-Jl,,; co e-Jl,,;
==L(i-l) +L.Ji ==L~+L-r-
;=0 i(i-1)(i-2)! ;=0 i(i-1)! ;=0(i-2)! ;=o(i-1)!
fJ3' = E( (X -oi) 3
= E(X ) = ~>3
00
;=0
-/1;
e .;
I.
oe . -JJ ;
= L{iU-l)(i-2)+3i(i-l)+i} e .;
;=0 l.
r 00
=e-JJfJ3L-fJ--+3e-JJfJ2L-fJ--+
;=o(i-3)!
;-3
i=0(i-2)!
00 ;-2 00
Li~
;=0
-JJ;
i!
t =e-'Il'(l+Il+~:
= e-JJfJ3eJJ+ 3e-JJfJ2eJJ+ fJ
+---00 )+3e-'1l2(1+1l+~:+---+E(X)
= fJ3 +3fJ2 + fJ
(iv) Kept beyond the scope of this book.
Theorem 2. If X has poisson distribution with parameter fJ then its
central moments are
= E(X) - X = fJ - fJ =0
(ii) Second central moment,
2
2 2
IJ2=E«X-X »)=E(X )-2XE(X)+X
, , 2
= fJ2 - 2 fJfJl + fJ = fJ 2 + fJ - 2fJ . fJ
'
+ fJ - = fJ
(iii) The third central moment,
.,.I fJ3 = fJ3' - 3fJ2'fJ/ + 2fJ/3 (using relation among central and moment)
pmj, fi = (7)pi(l-p r- i
, i=0,1,2, ...,n
fi of Poisson distribution.
distribution ==
L.
=;- L.
:. P(X = i) = fi
e-IJ . J.l.2 e-IJ. J.l.l
Now, f2 = t.. or, ----'-- = ----'-- :. J.I. =2
2! l!
:. P(X=l orO)=.t;+Jo=e-IJ(1+J.I.)=e-2(1+2)=3e-2
Mean of X = J.I. = 2
Now, 3rd moment about 0 = J.l.3'
_
-1- {e
-Il Ile-Il
+l!+~
112e -Il}
=1-e
-1- (
.0
()2
1+1.5+- -
1.5 J =0.19126·
2
Example 3. A radio active source emits on the average 2.5 particles per
second. Calculate the prob. that 2 or more particles will be emitted in an
interval of 4 seconds.
Here J.... = number of changes (which is particle emitted) per unit time
on an average = 2.5 .
Let X be the random variable denoting the number of particles emitted
in the given interval. Then X is poisson distributed with parameter 11 =
average number of particle in 4 seconds = 2.5 x 4 = 10 .
So the p.m.f, fi = P(X = i) = Ilie-Il / i! = 10ie-10 / i!
On an average there are 1 defective blade per 500 blades. So the average
number of defective blades in a packet of 10 is 10 x _1_ = ~ = 0.02.
500 50
Let X = number of defective blades in a packet. X follows poisson
distribution with parameter 11 = 0.02 .
e-Illli
So the prnf is P(X = i) = fi = -.,-.to
SPECIAL TYPE OF DISCRETE DlSTRmUTION I1S
1
.. Il = np = 200 x 50 = 4
Example 7. Six coins are tossed 6400 times. Using the poisson distribution
find the approximate probability of getting six heads 8 times.
Let X denote the number of six heads in the toss of six coin. Then X
EXERCISES
[T] SHORT ANSWER QUESTIONS
10. Prove that the probability of obtaining double six at least once in
1
24 throws with two dice is slightly less than 2" .
l1.Find the minimum number of times a die has to be thrown such
1
that the prob. of 'no. six' is less than 2".
[Hints: .The prob. of 'no. six' in n throws of a die =( !J
:. (5)
- < -1 :. n> -10g2 = 3.81]
6 2 10g5-10g6
2
12. A and B playa game in which A's chance of winning is "9. In a
series of 8 games, what is the chance that A will win at least 6 games ?
13. Find the probability that at most 5 defective pen will be found in
a box of 200 pen if experience shows that 2% of such pen are defective.
14. Five balls are drawn with replacement from a box containing 5
white and 4 black balls. Find the probability of getting 3 white balls.
[Hint. X = No. of white balls, X - b(n, p) with
n = 5, p = ~; P(X = 3) =?]
9
15. Two dice are thrown n times in succession. What is the probability
of obtaining double six at least once?
16. If probability of success is 0.1, how many trials are neccessary
1
in order that probability of at least one success is > 2" .
17. The mean of the poisson distribution is u , Then what is its
standard deviation? [ W.B.U.Tech 2004 ]
18. If a r.v. follows poisson distribution such that P{l) = P(2) , find
(i) mean of the distribution (ii) P(4) [W.B.u. Tech 2003,2004]
ANSWERS
7.B(5,:) 821
8. 3125
9. X: 0 1 2 3 4
1 3 27 27 81
.. t, : 256 64 128 64 256
.J3
Mean s j , s.d=T
there are exactly 2 defective screws in the lot (ii) no defective screw and
(iii) at most 2 defective screws.
1
4. The probability of a missile hitting a target is "4 .
(i) If 7 such missiles are sent, what is the probability of hitting
the target at least twice
120 ENGINEERING MATHEMATICS -IIA
.. Ii, = P(X = i) ( 3
=4 C\lO
)i ( 1- 3 )4-i =4
( 10
c,
3 )i ( 7 )4-i
10
10
4 3i x 74-i
= C·---
l 104
SPECIAL TYPE OF DISCRETE DlSTRmUTION 121
32 x 74-2
:. (i) P(X = 2) =4 C2· 104 = ·2646
4 x 74-0
4 Co' 3 10
(ii) P(X ~ 1) = 1- P(X = 0) = 1- to = 1- 4
74 2401
= 1- 104 = 1- 10000 = ·7599]
9. Show that the neccessary and sufficient condition for two given
numbers a and b to be respectively the mean and variance of some binomial
a2
distribution are that a> b > 0 and a _ b is an integer. Show further that
when these conditions are satisfied the binomial distribution is uniquely
determined.
10. Determine the expected number of boys in a family with 8 children,
assuming the sex distribution to be equally probable. What is the probability
that the expected number of boys does occur ?
P(X = 1), P(X s 1), P(X < 1), P(X > 1), P(1 ~ X ~ 3)
[Given e-2 =.1353]
122 ENGINEERING MATHEMATICS -IIA
16. (i) The average number of defective spot per sheet of a metal sheet
factory is 2. Asswning poisson distribution, what is the probability that a
particular sheet is free from spot? If the factory supplies 1000 such sheets,
how many sheet contain more than 2 spots ?
18. Find the probability that at most 5 defective bolts will be found in
a box of 200 bolts, if it is known that 2% of such bolts are expected to
be defective. (You may take the distribution to be poisson). Given
4
e- = 0.0183 .
j . 25. An office switch board receives phone calls at the rate of 2 in every
5 minutes on the average. What is the probability of getting exactly 4 calls
in 15 minutes ?
26. In turning out certain toys in a manufacturing process in a factory,
the average number of defectives is 10%. What is the probability of getting
exactly 3 defectives in a sample of 10 toys chosen at random, by using
Poisson approximation to the Binomial distribution (Take e = 2.72 )
(Hint: the method is adopted in an lllustrative Ex. page 89;
27. (i) An aeroplane runs with the help of its 1000 components. Failure
of a component is independent of other. Chance of failure of a component
'is 0.1 %. The plane takes off from one airport and reached safely its
destination if every component of its does not fail. Find the probability of
reaching safe. (Hint. Use method of approximation of Binomial variate by
poisson)
(ii) If 5% of the books bound at a certain bindery have defective
bindings, find the probability that 2 of 100 books bound by this bindery
will have defective bindings using (a) the binomial distribution (b) the
possion approximation [e-5 = 0.007].
124
ENGINEERING MATHEMATlCS-UA
ANSWERS
1. (a) 0.167 (b) 0.4073 (c) 0.4073 (d) 0.5920 2.0.105
6j(
3. (i) ( 2 50A
1 )2(49)4
50
49)6
(ii) ( 50 (iii)
(49)8
50
(6)(
+"1 1 )(49)5
50 50
4547
4. (i) 8192 (ii) 4 5. 11
27. (i) e-1 (ii) (a) 0.081 (b) 0.084 (iii) 0.080 (iv) 0.632
SPECIAL TYPE OF DISCRETE DISTRmUTION 12S
1. The mean of Binomial distribution B(n,p) (where n,p are the no.
of trials and probability of success) is
n
(a) p (b) 0 (c) np (d) I
[WB.U.T~ch 2007]
2. If X has Binomial distribution with parameter nand p, then its
variance is
(a) np (b) np 2 (c) np( 1 - p) (d) np( 1 + p)
3. The mean and variance of a Binomial distribution with parameter
2
6'3 are
4 4 2 4
(a) 6'3 (b) 4'3 (c) 0'3 (d) 9,-.
3
4. The mean of the Binomial distribution B( 10, i) is
P(X = 1) =?
4 16 32 16
(a) -
3 (b) 27 (c) 81 (d) 9.
7. The mean and standard deviation of a Binomial distribution are
11. A random variable has a poisson distribution such that P(l) = P(2) .
Then the S.D of X IS
(a) 0 (b) 2
ANSWERS
5.1. Introduction.
Let a discrete random variable X assumes 4 possible values
~, X ,
2
A1 and x4 • Corresponding to each of these vaules let another random
variable Y assumes 3 possible values YI' Y2 and Y3 .
Then the pair of variables (X, Y) assumes the 4 x 3 = 12 pair of values
shown below.
Y: YI Y2 Y3······ Yll
DISCRETE JOINT DISTRIBUTION & CHEBYSHEV'S INEQUALITY 129
Then the pair of variables or the bivariate (X, Y) assumes the values
shown in the following table
y
X Y2 Yn
X Y Yl Y2 Y3 Yn Total
.• Xl Pll P12 P\3 Pin PXI
EM-2A-9
ENGINEERING MATHEMATlCS-llA
130
m
m n n m
The grand total = LPXi =LPl'i = LLPij ==1
i=1 }=I }=I i=1
The above two tables represent the joint distribution of the bi-variate
(X, Y)
The row-wise totals PXi and the column-wise totals PYi are called
Marginal Probability Mass of X and Y respectively.
Pij = P(X = i, Y = Yj) are known as Joint Probability Mass of the
bivariate (X, Y)
Obviously for the individual variables X and Y,
P(X ==x) = PXi = Pi! + Pi2 + + Pin' j = 1,2,.·· =m
and P(Y ==Yi) = PYj = PI} + P2} + + Pm}' j = 1,2,.····n
:. X will have the probability distribution
X : XI X X3 •.••••• »; Total
2
(X, Y) is
Y : YI Y2 Y3 ... , Y n Total
discussed above.
If Pij =PXiPl'i i.e.,P(X=xpY=Yj)==P(X==xJP(Y==Yj) hold for
all values of j (1 s j s m) and j (l sj ~ n) then X and Yare called
independent random variables.
In the above Example the random variables X and Yare not independent
because we see P(X ==-1, Y = 8) ==0.5
but P(X =-l)P(Y ==8) = ·40x ·20 = ·08
DISCRETE JOINT DISTRIBUTION & CHEBYSHEV'S INEQUALITY 131
-3 5 7 Total
= L i
P Xi L
j
PYj': X, Yare independen
': . wherex.e A whereYjeB
= P(X E A) P( Y E B)
The proof of the converse part is left to the reader.
Example. Let X and Y be two discrete r~dom variables so that their pair
(X, Y) assumes the following bi-variate datas :
Y
3 5 8
X 3 5 8 Total
·40
-1 ·1 ·25 ·05
~- ------
DISCRETE JOINT DlSTRmUTlON & CHEBYSHEV'S INEQUALITY 133
_
2C 2 x5C I
5
_=_ I
- 10C 120 24
3
3C 1
P(3,0) = Probability of'3 red' = 12~ = 120
x -
0
1
1
1
2
1
Total
7
0 - - -
12 6 24 24
1 1 1 21
1 - - - -
4 4 40 40
1 1 7
2 - - 0 -
8 20 40
1 1
3 - 0 0 -
120 120
7 7 1
Total - - -
15 15 15 1
In fact above two tables persent the joint distribution.
Now, P(X s 2, Y ~ 1)
= P(2,1) +P(2,2) +P(l,l)+P(1,2) +P(O,l) +P(0,2)
1 1 1 1 1 8
=-+0+-+-+-+-=-
20 4 40 6 24 15
The marginal distribution of Y is given by
Y : 0 1 2
7 7 1
PYj :.
15 15 15
Probability of "more than 1 white balls"
7 2 9 3
P(Y ~1)=-+-=-=-
15 15 15 5
1 7 7
From the above table we see 12 * 24 x 15
:. X, Yare not independent.
5.4. The Multinomial Distribution
We can define joint probability distribution for n number of random
variables in exactly the same manner as we did for 2 random variables.
One such joint distribution is multinomial distribution.
DISCRETE JOINT DISTRIBUTION & CHEBYSHEV'S INEQUALITY 135
The joint distribution of (XI' X2,· ••• X r) given by this joint probability
is called Multinomial Distribution.
Note that when r = 2 the multinomial distribution is reduced to binomial
distribution where SI = success, S2:::: failure.
Example. Suppose that an unbiased die is rolled 9 times. Here E is the
experiment of rolling the die. So we are considering 9 independent trials
of E. E has 6 possible outcomes '1', '2', .... '6' with respective
pro b a biliti
1 ities P ::::-1 P = -1 .... P = -1 .
1 6' 2 6' 6 6
Let XI :::: Number of 'face l' occured in 9 trials.
X2 = Number of 'face 2' occured in 9 trials.
9'
= . .- 1 [v here L:n. =9]
n !n2 !... n6! 69 I
9! 1 9! 1
3!2!2!1!1!0! 69 = 3!2!2! 69 •
_ 21 7 49
-PXl +Pxo =-+-=-
40 24 60
DISCRETE JOINT DlSTRffiUTION & CHEBYSHEV'S INEQUALITY 137
Theorem. For any two independent random variables X and Y the joint
distribution function F (x, y) is F (x, y) = Fx (x) Fy (y).
Proof. The proof follows from the previous theorem.
5.6. Sum of independent Random Variables.
In this section we discuss the behaviour of the sum X + Y where X
and Yare independent random variables having well known distribution.
Theorem 1. If X and Yare two independent random variables having
binomial distribution with parameter (m, p) and (n, p) respectively then
their sum X + Y has binomial distribution with parameters (m + n, p)
Proof. By hypothesis the distribution of X is
X. 0 1 2········m
/; fo J; 12 r;
where /; = P(X = i) = mCipi qni-i and the distribution of Y is
Y: 0 1 2 ········n
The event (Z = i) = (X + Y = i)
=(X = 0, Y"=i)U(X =1, Y =i -1)U ····U(X = i, Y =0)
:. P(Z = i) =P(X = 0, Y = i) + P(X = 1, Y = i-I) +> ··+P(X = i,Y = 0)
= P(X = e}P( Y = i) + P(X = I)P( Y = i-I) + ... + P(X = i)P(Y = 0)
.. X and Yare independent
i
= 'Lfrgi-r = 'L mCrpr
i
r=c;»: q"-(i-r)
r=O r=O
t
= '"
L... piqm+lI-i In C
r
II c.i -r
r=O
i
= piqm+lI-i '" mC II c.t-rr
L..J r (1)
r=O
138 ENGINEERING MATHEMATICS -IIA
(n Co + n CJX-t n
C2X
2
+ ... + n CiX inn)
+ ... + Cnx (2)
On RHS of (2) Coefficient of Xi is
i
"" "c r
== L..J "C.t+r
r=O
r=O
From (1) we get P(Z == i) == m+nCi/qm+lI-i
:. Z i.e., .k + Y has binomial distribution with parameter m + nand p.
Theorem 2. If X and Yare two independent random variables having
poisson distribution with parameters PJ and P2 respectively then their
sum, X + Y has poisson distribution with parameters PJ + P2 .
Proof. The values of X and Yare respectively
X:01234
Y:01234
Let U=X+Y.
:.U : 0 1 2 3 .
Now, for any integer k > 0
P(U=k)=>LP(X+Y=k)~ L P(X=i,Y=})
i+j=k
"
-PI
~.
i
e
-P2
fJ.2
j J =e-(PI+P2) ""
I
PIfJ.2
j
e -(J11 + J12)
= O! (f.l2 + f.l1)O.
Thus For any integer k ~ 0
e -(J11 + J12) (f.l + f.l )k
P(U=k)= 1 2
k!
e-(J1I+J12) (f.l + f.l )k
or, P(X + Y = k) = . I 2
k!
.'. X + Y has Posisson distribution with parameter f.ll + f.l2
5.7. Bivariate Expectation
Let (X, Y)be a bivariate having joint probability mass
P ij = P( X = Xi Y = Yj ) , marginal probability masses P Xi and P tj
j .
E(~(Y») = L~(Yj}PYj
j
140 ENGINEERING MATHEMATICS - IIA
If the concerned series are infinite series then the expectations exists
provided the infinite series is absolutely convergent.
Example. For the joint distribution of (X, V),
X -1 2 3 Total
X YI = -1 Y2 = 2 Y3 = 3 Total
3
E(Y + 2) = L (yj + 2)pYj = (YI + 2)pYl + (Y2 + 2)prz + (Y3 + 2)Pn
j=1
= LLX;Pij ± LLYjPij
;=1i=) ;=1 j=1
In n n n
=LX;LPij±LLYjPij
;=1 j=1 j=1;=1
m n m
= LX;Px; ± LYjLPij by definition of Px;
i=1 j=) i=1
m n
= LXiPXi ± LYjPYj by definition of PYj
;=1 j=1
=E(X)±E(Y)
Note. This theorem can be extended to more than two variables e.g
E(X + Y + Z + W) = E(X)+E(Y)+ E(Z) + E(W)
Theorem 2. For a joint distribution of (X, Y)
E(XY) = E(X)E(Y) provided X and Yare independent random
variables.
m n
Proof. E(XY) = LLx;YjPij
i=lj=1
m n
= LLxiYjPXiPYj" .: X and Yare independent
;=lj=1
m n
= LXiPXi LYjPYj
i=1 j=1
=E(X)E(Y)
Note. Converse of this theorem is not true . See a subsequent Example.
l-- --J
142
ENGINEERING MATHEMATICS -IIA
Theorem 3. E(k¢(X,Y»=kE(¢(X,Y»
Proof. Left to the readers.
x 0 1
.
2
1 0·2 0·2 0
X Yl == 0 Y2 == 1 Y3 == 2 Total
J
144 ENGINEERING MATHEMATICS-llA
2 3
Now, E(XY) == L,L,xiYjPij
i=1 j=1
2
== L, (XiYIPiI + XiY2Pi2 + XiY3Pi3)
i=1
2 2 2
== YIL,XiPiI + Y2L,XiPi2 + Y3L,XiPi3
i=1 i=1 i=1
2
E(X)== L,XiPXi ==XIPXI +X2PX2 ==OxO·6+1xO·4==0·4
i=1
. 3
E(Y) ==L,YjPYj ==YIPYI + Y2Py'2 + Y3Pn
j=1
==OxO.3+1xO.6+2xO.l ==0·6+0·2==0·8
and CJ
y
==JE(y )-{E(y))2
2
Similarly a; =Ici ay
Again Cov(U,V) = E{(U -U)(V -V)}·
=E{(aX +b -aX +b)(eY + d -eY + d)}
Cov(V,V) ae Cov(X,Y) ae
Then P = = =--P
uv 0" uO"v lalO"x lei 0" y lallel xy
Property 4. If V = aX + b, V = ey + d then
(i) Puv = P xy if a and e are of same sign
(ii) Pliv = -Pxy if a and e are of opposite sign.
ae ae
Proof. By previous property Puv = lalicl Pxy = lael Pxy
(i) Let a and e are of same sign. Then ae > 0
ae
:.Iael = ae. Then PI/V = =P» = Pxy.
LIe
(ii) Let a and e are of opposite sign. Then ae < 0
ae
:.Iael = -ae. Then Puv = -Pxy
-ae
= -Pxy
Property 5. -1 ~ Pxy s1
Proof. We construct two variables V and V by the rule
V = X - X and V = Y - Y
o"x O"y _
Cov(X,Y)
= = Pxy
O"xO"y
DISCRETE JOINT DISTRIBUTION & CHEBYSHEV'S INEQUALITY 147
Again E(U2)=E{(
lX ;xx) Y} =~E{(X
o"x
-xi} =~O";
o"x
=1
Similarly E(V2)=1
Now, (U + vf ~0
or, U2 + 2UV + V2' ~ 0
2
:. E (U + 2UV + V2) ~ 0
Now, U=
- (x -)( J. A-:
--
O"x
=--=0
0"x
Similarly, V=0.
:. Cov(U,V) = E( (U - U)CV - V)) = E(U V) (3)
or, E(UV)::;~=l
2
or, COy (U ,V) ::;1 using (3)
X 4
1
1 0 1 4 ,
PXi
- 0 0 0 0 1/5
-2 5
1
- 0 0 1/5
-1 0 0
5
1
0 0 0 - 0 0 1/5
5
1
1 0 0 0 - 0 1/5
5
1
2 0 0 0 0 - 1/5
" 5
1 1 1 1 1
PYj - - - - - 1
5 5 5 5 5
5 5
Here E(XY) = LLxiYjPij
i=1j=1
5
= :~:>i(YIPii + Y2Pi2 + ..... + YSPiS)
i=1
.............................................
1
+XS(YI xO+ Y2 xO+···+ Ys x-)
5
DISCRETE JOINT DISTRIBUTION & CHEBYSHEV'S INEQUALITY 149
1
= -(XIYI + X2Y2 + ... + xsYs)
5
1
= -(-2 x 4+ (-I)x 1+ Ox0+ 1xl + 2 x4)
5
1
=-(-8-1+ 1+8) = 0
5 .
s 1 1 1 1 1
E(X) = LX;Px; = -2 x-+ (-I)x - +Ox -+ 1x -+ 2x- = 0
;=1 5 5 5 5 5
s 1 1 1 1 1
E(Y) = L.YjPYj :;::::
4x-+ I x-+ Ox-+ 1x-+4x- =2
j=1- 5 5 5 5 5
Thus Cov(X, Y) = E(XY) - E(x)e(y) = 0 - 0x 2 = 0
Proof. Let Z =X - Y
:. E(Z) = E(X - Y) = E(X)
- E(Y)
E(Z2) = E( (X - Y)2) = E(X2 - 2XY + y2)
More generally,
Var(alX1 +a2X2 + .... :+anXn)=
"
~>iVar(X;)
;=1
+2L,L,a;ajCov(X;,Xj)
; <i
DISCRETE JOINT DlSTRffiUTlON & CHEBYSHEV'S INEQUALITY 151
1
=--zVar(XI +X2 + .... +Xn)
n
2
1 2 2 2 1 2 a
=-(a +a +.... ·+a )=-·na =-
n2 n2 n
3. 6 PXi
1
-
5 PII PI2 2
1
-
7 P21 P22 2
1 2
PYj
- -
3 3
1
Let PII =x. :. PI2 =2"-x
P21 =.!..- x and P22 =~- Pl2 = ~-(.!.. -x) =.!..+ x
3 3 3 2 6
1 1
Now, E(X) =5x-+ 7x-= 6
2 2
1 2
E(Y) = 3x-+6 x- = 5
3 3
E(XY) = 5 x 3x Pll + 5x 6x Pl2 + 7x 3x P21 + 7x 6x P22
3 6 Total
(Pxi)
1 5 1
- - -
5 12 12 2
I 1 1
- - -
7 4 4 2
1 2
Total - - 1
3 3
(PYj)
DISCRETE JOINT DISTRIBUTION & CHEBYSHEV'S INEQUALITY 153
1 1 1 1
(ii) We see PI I =- but P Xl X PYI =- x- =-
12 2 3 6
:. Pu "# PXI . PYI
o 2
>( 0 1 2 PXi
X 2
Y o 1 2
PYj 0 .4 O· 2 O· 4
0 1 2 Total
~x·
1 P"x,y, =
·3xO ·2xl ·lx2 0·4
2 -l x 0 ·Ox2 ·3x4 1· 2
.', E(XY) = 1· 6
Example 3. The bivariate (X, Y) is such that it take the values (i,j)
where i = 0,1,2,3;j = 1,2,3,4. The joint probabilities P(X = i,Y = j)
= k(3i + 4j). Find
(i) the value of k
(ii) the marginal distribution of X and Y.
(iii) P(X ~ 2, y s 3)
(iv) P(Y = 2/ X,,; 3)
(v) Are X and Y independent?
3 4 3 4
Now, 2:2:Pr =1 or, 2:2:k(3i+4j) =1
IJ • 0 . 1
i=O)=1 1= J=
3 4 3 4 4
or, k2:2:(3i+4j)=1 or, k2:(2:3i+ 2:4j)=1
i=O)=1 i=O )=1 )=1
3
or, k2:(3ix4+4(1+2+3+4»=1
i=O
3
or, k2:(12i+40)=1
i=O
or, k ( 12 t t i+ 40) =1
or, k(12(1+2+3)+4x40)=1
1
or, 232k =1 :. k =-
232
(ii) Now, X 0 1 2 3
444
Now, ix(X =0)= 2: PO) = 2:k(3xO+4j) =4k2:j
)=1 )=1 )=1
. 40 5
= 4k(1+ 2+ 3+4)= 40k = -=-
232 29
4 4 .
ix(X =1) = 2: PI) = 2:k(3xl+4j)=k(4x3+4x(1+2+3+4»
)=1 )=1
=52k=~=Q
232 58
156 ENGINEERING MATHEMATICS -IIA
8 19
Similarly we shall find Ix (X = 2) = 29' Ix (X = 3) = 58
.'. the marginal distribution of X is
X 0 1 2 3
5 13 8 19
Ix - - -
29 58 29 58
Similarly we can find the marginal distribution of Y is (detail evaluation
is not shown)
Y: 1 2 3 4
17 25 33 41
Iy :
116 116 116 116
(iii) P(X~2,Y::;3)=:L:LPij
i:?2jS3
3 3
= :L:L k(3i + 4 j)
i=2j=1
= k~
3(3 f;3i +4f;/
3 J 3
= k~(9i +24)
( 3)
:::.k 9:Li + 24x 2 = k(9(2+ 3)+ 48)= 93k =-
~2
93
232
13 25 325
58x116= 6728
x
but PXl px2=
.'. X and Yare not independent
DISCRETE JOINT DlSTRlBUTION & CHEBYSHEV'S INEQUALITY 157
x2
0
-05 -10
2
-25 0-4
Solution. o 2
X2 0
For i = 0,1,2 and j = 0,1 ,
J'----'---- __ -.JL
158 ENGINEERING MATHEMATICS-11A
o 1
POO Po,
2
4 2C C 4 2 3
lXI' = PIO + Pll =10+ 5~2 0 = 10 + 10 ="5
1
=
10
:. the marginal distribution of X, is
o 2
3 3 1
10 5 10
DISCRETE JOINT DISTRIBUTION & CHEBYSHEV'S INEQUALITY 159
X 2
1
3
1
4
1
PXi
2
1 - - - -
4 4 6 3
1 1 1 1
2 - - - -
12 12 6 3
1 1 1
PYj
- - -
1
3 3 3
1 1 1
P(X=l Y=3)=---=-
, 3 12 4
1 2
Px(X=l)=l-3"=3"
2 3 4 Total
~x·
1 23
1 PllX1Yl= PI2XIY2=
lx4x- -
6 12
1 1 1 3 2
-xlx2=- -xlx3=- -
4 2 4 4 3
1 1 1 1 1 4 13
2x2x-=- 2x3x-=- 2x4x-=- -
2 12 3 12 2 6 3 6
49
E(XY)= -
12
49 4 1
Cov(X,Y) = E(XY) - E(X)E(Y) = ---x 3 =-
12 3 12
:. the correlation coefficient,
1 1
Cov(X,Y) 12 12·.fj
P,,= J.J, =J%~= 31 =s
(ii) When X = 1, Y = 2, X + Y = 3; when X = 1, Y = 3, X + Y = 4.
Thus
X +Y: 3 4 5 6
1
P(X + Y = 3) = P{(1,2)} = 4
1 1 1
P(X + Y = 4) = P{(1,3),(2,2)} =4+12 ="3
1 1 1
P(X + Y = 5) = P{(1,4),(2,3)} ="6+ 12 = 4
1
P(X +Y = 6) = P{(2,4)} ="6
DISCRETE JOINT DISTRlBUTION & CHEBYSHEV'S INEQUALITY 161
X+Y 345 6
1 1 1 1
4 3 4 6
Example 7. If X and Y have same standard deviation then prove that
X + Y and X - yare uncorrelated.
Solution. Let U = X + Y and V =X - Y
:. U = X + Y and V = X - Y
7!
=2!1!0!2!0!2!"6
( 1 )2 ( "61)1( "61)0( "61)2( "61)0 ( "61)2
according to Multinomial Distribution.
7! 1 630 35
="8' 6' = 279936 = 15552
Example 9. An urn contains 5 white, 2 blue and 3 red balls. Six balls are
drawn with repalcement. Find, using multinomial distribution, the
probability that there are four white' and two red balls.
Solution. Let E be the experiment of drawing a ball from the urn. This
has three possible outcomes: white ball (w), blue ball (b) and red ball
(r) with respective probability
5 1 2 1 3
PI = 10 =Z,P2 = 10 =S,P3 = 10
Since six balls are drawn with replacement we think E is repeated 6
times.
Let XI = number of W, X2 = number of band X3 = number of r in
6 trials.
By Multinomial distribution, required probability
=p ( XI =4'X2 =0,X3
6!
=2 ) = 4!0!2!PI
4 °
P2 P3
2
X3 = No.
of occurence of S3 among n = 6 trials.
From the principle of multinomial distribution,
Required probability = P(XI = 2,X2 = 3,X3 = 1)
./
6! 2 3 I 720 ( 1 )2 ,( 1)3 (, 1 )1
= 2! 3! 1!PI P2 P3 = 12"4 "2 "4
60 15
=-=-
512 128 "
Example 12. XI'X2,X3 are three discrete random-variables which are
pairwise uncorrelated. Each has same standard deviation. Find the
correlation coefficient between XI + X2 and X2 + X3 .
{E(X2)+E(X3)}
164 ENGINEERING MATHEMATICS -11A
. Cov(U, V) 0'2 1
• '.
p=
uv <«. = .J20" JiO' =-2
Example 13. Find the expectation of the total number of points in a bridge
hand of 13 cards where the points are assigned as 2 for spades, 3 for
club, 4 for heart and 6 for diamond.
.Solution, Let X\ = points scored in the first card received
Then X\ 2 3 4 6
1 1 1
Prob:
4 4 4 4
I 1 1 1 15
:. E(X\) =2x-+3x-+4x-+6x- =-
4 4- 4 4 4
If X2,X3 .. · .. X\3 be the points scored in the second, third,
thirteenth card respectively then as above we shall get
E(X )=~
E(X )=~ .. ·..E(X )=~
4 '
2 3 4' \3 4
:. -the expectation of the total number of points
= E(X\ +X2 +X3 + .... +X13) =E(X\)+E(X2)+ .. ·· .. ·+E(X13)
= 13x~ = 195
4 4
Example 14. If X and Yare two independent random variables having
standard deviation 0'\ and 0'2 prove that the correlation coefficient of X
a)
and X + Y is I 2 2
'Va) + 0'2
DISCRETE JOINT DISTRIBUTION & CHEBYSHEV'S INEQUALITY 165
Solution. Let U =X +Y
Now, U = E(U) = E(X + Y) = E(X) + E(Y) =X +Y
Cov(X,U') = E {(X - X)(U - U)}
=E{(X -X)(X +Y -x -f)}
=£(X -X){(X -X)+(Y -y)}) .
=E{(X _X)2 + (X -X)(Y - Y)}
=E(X -X)2)+E(X -X)(Y -Y»)
= Var(X) +COV(X,Y)
=0"2
1
+0"22
2
. _ Cov(X,U) _ 0"1 _ 0"1
.. Px •u - 0"0"
- ~ 2
or, Px ,x+y - ~ .
X U 0"1 0"1 + 0"22 2
0"1 + 0"22
Example 15. Two discrete random variables X and Yare connected by
the relation 2X + 3Y + 4 = 0 . Prove that the correlation coefficient between
X andYis -1.
Solution. Given 2X + 3Y + 4 = 0 (1)
or, 2X + 3Y + 4 = 0 (2)
166 . ENGINEERING MATHEMATlCS-IIA
= E {(X - X)( - 2X + 4 + 2X + 4
3 3
J} using (1), (2)
2 2
--a
3 x
=-1
2
ax ·-a
3 x
Example 16. 'Player A tosses 5 coins and player B tosses 8 coins. If the
coins are unbiased, find the probability of obtaining a total of 6 heads by
the two players.
Solution. Let X = number of heads among 5 trials by A
Y = number of heads among 8 trials by B
:. X + Y = total number of heads obtained by A and B.
m +n = 5 + 8 = 13 and p = -1
2
Required probability = P(X + Y = 6)
Example 17. An urn contains four white, four red and four black balls.
The balls of each colour are marked 0, 1, 2, 3 respectively. A ball is drawn
from the urn at random. A random variable X assumes the value 0, 1, 2
according as the ball is white, red and black respectively. Y denote the
number marked on the ball.Find E(X), E(l'). Hence find the covariance
of the variates.
Solution. X o 2
Y o 2 3
Now, Poo = P(X = 0, Y = 0) = Probability that the ball is white with
marking 0 =~
12
1 1 \-I ••
Similar9" POI = P02 = P03 = 12 . In general Pi} = 12 v t.]
x 0
1
1
I
2
1
3
1
Total
1
0 - - - - -
12 12 12 3
1 1 1 1 1
- - - - -
1 12 12 12 12 3
1 1 1 1 1
2 - - - - -
12 12 12 12 3
1 1 1 1
Total - - - -
4 4 4 4
Marginal distribution of X is
X 0 1 2
111
Ix; - - -
3 3 3
1 1 1 1
:. R(X) =Ox-+ 1x-+ 2x- = 3x- =1
3 3 3 3
That of Yis
Y o 2 3
1 1 1 1
4 4 4. 4
168
ENGINEERING MATHEMATICS -IIA
1 1 1 1 6 3
:. E(Y) = Ox-+ lx-+2x-+3x-=_ =_
4 4 4 4 4 2
From the above table we see Pij = Ix; x IYj V i,j
:. X and' Yare independent.
:. Cov(X,Y) = O.
Example 18. Two balls are drawn one by one without replacement from
a box containing three balls, numbered 1, 2, 3. Let
X = number on the first ball drawn
Y = number which is larger of the two.
(i) Find the joint distribution of (X, Y).
(ii) Cov (X, Y)
Ie x le 1
PI2 =P(X=I,Y=2) =P(I,2) = 3 I 2 I =_
CjxCI6
Ie xl e 1
PI3 = P(X = 1,Y = 3) = P(l, 3) = 3 I 2 ~ =6
CI X 'JI
P23 =P(X=2,Y=3)=P(2,3)=
'c
3 I
xl
2
e1=6 1
Clx CI
P32 = P(X = 3,Y = 2)
-
DISCRETE JOINT DISTRIBUTION & CHEBYSHEV'S INEQUALITY 169
X 2
1
3
1
Total
1
(lxJ
1 - - --
6 6 3
1 1 1
2 - - -
6 6 3
1 1
3 0 - -
3 3
1 2
Total (fYj - - 1
3 3
(ii) To find E(XY) = LLPij i- j we construct the following table of
(Pi) . i- j) .
<; 2 3 Total
E(X) = lx-+1112x-+3x- =2
. 3 3 J
E(Y) = 2 x.!. + 3 x 3. = ~
3 3 3
11 8 1
:. COY (XY) = E(XY) - E(X)E(Y) = 2" - 2 x 3' = '6
170 ENGINEERING MATHEMATICS-IIA
13 .
DISCREET JOINT DISTRIBUTION & CHEBYSHEV'S INEQUALITY 171
Illustration.
Let X be the number of Chips produced in a factory during a week. Its
mean and variance be 50 and 25 respectively. Find the probability that this
week's production will be between 40 and 60.
Now, 40<X <60
or,40<X<60<=:>IX-501<10 (1)
By Chebyshev's inequality,
1 . 1 1
P {IX -5ol~10 } ~-2 Var(X) =-x25=-
10 100 4
172 ENGINEERING MATHEMATlCS-1IA
or,l-p{IX-501<1O}:::;!..
4
or, P {IX - 501< 10} ~ 1-!.. = l
4 4
or, P(40 < X < 60)-~ l using (1)
4
:. the required probability is at least l
4
From the above theorem another result known as One-sided
Chebyshev's inequality can be deduced. This is stated in the follwing
theorem.
Theorem. If X is a random variable with mean 0 then for arbitrary E> 0
Proof. The proof is kept beyond and the scope of the book.
Corollary. If the mean of X is 0 then for arbitrary E> 0 .
P(X:::;-E):::; Var(X)
Var(X)+ E2
Illustrative Examples.
Example 1. The distrubution of a random variable X is given by
1 3 1
P(X =-1) =8'P(X = 0) =4"'P(X = 1)=8'
1
pOX -0\ ~ e )~4
e
Now we consider the following cases:
Case (i) Let 0 < e ~ 1
p{IX-J1I,?~}$ ~ Var(X)
n2
or, lim
n-+oo
p{IX - J11 '?~}=
n
lirn 0 = 0
n-+oo
DISCREET JOINT DISTRIBUTION & CHEBYSHEV'S INEQUALITY 175
X-,u*-O]
= 0 if mL is paired with mJ
Similarly consider X 2 such that
X = 1 if m2 is paired with some Wj
2
=0 if m2 is paired with some m],
100 100 99
Thus P(XL = 1)=- :. P(X( =0)= 1--=-"
199 199 199
:. the distribution of XI is
XI o 1
99 100
Prob:
199 199
176
ENGINEERING MATHEMATlCS_11A
Now, Pi X, = 1) = _100
I 199
Now if m; is paired with some wj,mj can be paired with remaining
99
Now, £eX)=Ox-+lx_=_ 100 100
199
I 199 199
., £(X) = £(X1) + £eX2) + ... + £(XIOO)
= LVar(XJ+2LLCoV(X;,Xj)
;=1 i< j
= 100(1_100)= 100 x 99
199 199 199 199
If i < j,Cov(Xj>X;> = E(Xi,Xj) - E(Xi)E(Xj)
100 99 100 100
=-x---x-
199 197 199 199
100 (99 100)
= 199 197 - 199
100 1
=-x--
199 39203
100 100 99 100 1
Thus Var(X) = I-x-+ 2x
II-x--
j=1 199 199 199 39203
i<j
=100x
100 x 99 -!-2x
100C x 100 x_l_
199 199 2 199 39203
= 25 ·126 (Approx).
X = 30 ~ Z = 30 - 50·25 = -20·25
Now, X s 30 ~ Z s -20·25.
:. by One sided Chebyshev's inequality (in fact by the corollary)
P(X s 30) = P(Z s -20·25)
< Var(Z) = Var(X)
- Var(Z) + (20· 25)2 Var(X) + (20.25)2
25 ·126
= = ·058 (approx).
25 ·126 + (20·25)2
:. there is a maximum chance of ·058 i.e., 5.8% of the required event
Example 5. The average number of bikes sold weekly at. a certain show
1
room is 16 with s.d 8" . Find the maximum probability that next week's sale
will exceed 18.
Solution. Let X = number of bikes sold weekly.
By problem X = 16 . Let Z = X -16 . Then Z=0 .
1)2 1
Var(Z) = Var(X) = ( 8" = 64 .
EM-2A-12
178 ENGINEERING MATHEMATICS - IIA
8
or, p(IX - yl > 15) ~ ;2 5 - p(lX - YI = 15)
28
~ 225 (X may be equal to Y)
7 1 1
and c: = Var(X) = np(l- p) = 2000 x-x - = 500.
2 2
Now P(900 < X < 1100) = P(-IOO < X -1000 < 100)
I 19
:. from (1) P(900<X<1l00)~I--=-
20 20
So the probability that the number of heads lies between 900 and 1100
19
is at least 20'
Example 8. The p.m.fofa random variable Xis I(X = 1) = Ti,i = 1,2, .. ·..
Show that Chebycheff's inequality gives pOX - 21~ 2) ~ i.Also fmd the
actual probability.
l+x 2 2 2
:. = 1+ 2 + 3 x + ... (2)
(1- x)3
-
~----~------~~------------~
180 ENGINEERING MATHEMATICS-ItA
1 )-2 =1+2'2+13'"22+'"1
( 1-2
1 1 1 1
4·-=-+2·-+3·-+···
2 2 22 23
1 1 1
:.-+2·-+3·-+···= 3 2 :. m =E(X)= 2
2 22 2
0-
2 = Var (X) =E(X2)-{E(X)}2 =6-4=2
Cov (X,Y)=-3, fmd the maximwn probability of the event (Y > X + 15)
Solution. Let Z = X - Y :. Z =X - Y =0
Var(Z) = Var(X - Y) = 10+ 12-2x-3 = 28.
P(Z$-15)$ Var(Z) = 28 28
Var(Z) + 152 28 + 225 253
. 28
or, P(X-Y$15)::;-
253
. 28
or, P(Y~X +15)::;-
253
28
or, P(Y > X +15)+P(Y =X +15) ~-
253
28 28
or, P(Y >X -15)$--P(Y =X +15) ~-
253 253
28
:. the required maximum probability is 253 .
Example 10. From the past experience, a professor knows that the test
score of a student taking her final examination is a random variable with
mean 75 and variance 25. What can be said about the probability that a
student will score between 65 and 85 ?
Solution. X = score of the student
:. X = 75, Var(X) = 25.
Let Z = X - 75 . So X = 65 ~ Z = -10 and X = 85 ~ Z = 10
:. P(65 < X < 85) = P(-1O < Z < 10) = p(IZI < 10) (1)
Now, Z=X -75 = 75 - 75 = 0 and Var(Z) = Var(X) = 25 .
By Chebychev's inequality we have
V.
or, p( 65 < X < 85) ~ ~ using (1)
4
:. The probability of the event is at least
') .
182 ENGINEERING MATHEMATICS - DA
20
Solution. Let X = LX; . Since X; are mutualy independent soXis poisson
;=1
Now, P(X > 15) = P(Z + 20 > 15) = P(Z > -5) (1)
By Chebyshev;s inequality
Distribution of X; is
1 2 3 4 5 6
1 1 I 1 1 1
6 6 6 6 6 6'
1 21 7
:. E(X) = -(1 + 2+ 3+4+5 +6) = - =-
. I 6 6 2
Var(X;)=E(X;2)_{E(X;)}2·
Exercise 5
1. Following is' the joint distribution of X and Y.
X 1 2 3
1 3 1
0 - - -
10 10 10
1 1 1
2
- - -
5 10 5
(iii) X and Y
(iv) O'x,O'y and Pxy
7. An urn contains 2 white, 3 black and 2 red balls. Three balls are
drawn one by one without replacement. Let.x: denote the number of white
balls, Y denotes number of black balls. Then obtain the joint distribution of
the two variables X and Y.
9. An urn contains three balls marked with numbers 0, 1,2 and having
white, red and black clours respectively. One ball is drawn at random. Let
X takess value 1, 2 and 3 according as the colour of the ball is white, red
and black. Prove that the correlation coefficient between X and the number
marked on ball is 1.
P(X = 10) =.!.. Y can take the values 5, 6, 7 with the probability
3
P(Y = 5) = .!.,P(Y = 6) =.!.. Find expectation of Z.
4 2
15. Th~ joint distribution of (X, Y) is given by P(X = O,Y = 0) =~ .•
1 I
P(X = I,Y = 0) ="4'P(X = -l,Y = 0) = "4'P(X = O,Y = 1) ="4I show that
X and Yare uncorrelated but X and Yare not independent.
16. Let X assumes the values ±l,± 2 each with probability .!..
4
Let Y = X2 . Construct the joint distribution of (X, Y). Show Pxy =0.
Are they independent?
I -1 2 -2
Y
1 1/4 1/4 o
4 o o 1/4
17. A box contains 5 white, 2 blue and 3 red balls. If 6 balls are drawn
one by one with replacement find, using multinomial distribution, the
probability that there are 0) 2 white, 3 blue and 1 red balls (ii) 2 balls of
each colour.
18. Find the probability that in 8 throws of a die, the number 1, 3, 5 turn
up 2, 3,.3 times respectively
ENGINEERING MATHEMATICS -UA
186
[Hint:
10!
4! 2!(1 !)4
( 1
6"
)10 by multinomial distribution]
20. A die is rolled 8 times. What is the probability that the faces 1,3,5
turn up 2, 3, 3 times respectively.
21. A box contains 10 balls of which 5 are white, 3 red and 2 black. A
ball is drawn and replaced 3 times. Find the probability that the balls are of
different colours.
22. A pack of 8 cards contains 3 aces, 2 kings, 2 queens and 1 jack. The
pack is shuffled 5 times and the top card is exposed each time. Find the
probability that the exposed cards are ace twice,. king once, queen once
and jack once.
.
[Hmt:
5! (3
2! 1!1!1!"8
)2 (2)1
"8
(2)1
"8
( "8]1)1
23. A box A contains six balls numbered 1, 2, 3, 4, 5, 6 and another box
B contains six balls numbered 1, 2, 3, 0, 0, 0. Two balls are drawn at
random, one from each box. Let X = No. of the ball drawn from box A
and Y = No. of the ball drawn from box B. Find E(XY).
[Hint: Obviously X and Yare independent. B(Y) = "I. .!. + 2 . .!. + 3 . .!.]
6 6 6
24. Three balls are placed into three cells at random. Let N denote the
number of occupied cells. Xi denote the number of balls in r-th cell (for
i = 1,2,3). Find the joint distribution of (X\,N) and of (X\,X2) . Hence
find Cov(N,XI) and Cov(X\,X2)·
25. Two unbiased coins are tossed.
X = 1 if first coin shows head
= ° if first coin shows tail.
Y =1 if second coin shows head
= 0 if second coin shows tail
(i) Find the joint distribution of (X, Y).
(ii) Marginal distribution of X and Y.
(iti) Are X and Y independent. (iv) Find E(2X + 3Y)
DISCREET JOINT DlSTRmUTION & CHEBYSHEV'S INEQUALITY 187
27. The bivariate (X, Y)takes the value (i,}) where i = 1,2; 3;} = 1,2,3.
P(X = i, Y = }) = kij where k is constant. Find
(i) the value of k (ii) P(l ~ X s 3,Y ~ 2)
= Ys~(i,})= (2,2)
Find expectation 0'[ (i) X (ii) Y (iii) X +Y
and (iv) Cov(X,Y) (v) Var(X +Y)
29. In a joint distribution of the two variables X, Y if ax' ay, al._y are
standard deviation of X, Yand X - Y respectively then prove that the
2 2 2
ax +ay +ax_y
correlation coefficient Pxy = 2
axO"y
30. The joint distribution of (X, Y) is given by
o
o a e
b d
34. If X = f.1 and Var(X) = a2 then for arbitrary E> 0 prove that
a2
P(X > f.1+ E) :5; 2 2'
a +E
35. The average number of automobiles sold in a week in a showroom
is 16 with variance 9. Find the least probability that next week's sales are
between 10 and 22 inclusively.
40. From the past experience a professor knows that the test score of a
student taking her final examination is a random variable with mean 75 and
variance 25. Find the maximum chance that a particular student will score
more than 85.
41. Show that the probability that the number of heads in 2000 throws
with a fair coin lies between 900 and 1100 is 0.999. Compare this value
with the value given by Chebycheff's inequality.
42. A symmetric die is thrown 360 times. Determine the lower bound
for the probability of getting 50 to 70 ones using Chebycheff's inequality.
43. The average number of cars sold weekly at a certain dealership is
16 and the s.d is 3. Find a lower bound to the probability that next week;s
sales are between 10 and 22 inclusuvely.
44. Does there exists a variate X for which
P(p - 2~ 5, X 5,P + 2~) = 0 .6 ?
[Hint: the given probability contradicts Chebycheff's inequality]
45. If X is the number of points on a dice prove that Chebycheff's
inequality gives p(IX - Xl> 2.5) < 0·47 when the actual probability is
nearly O.
46. If X be a r. v such that E(X) = 3 and E(X2) = 13 then using
50. Use Chebycheffs inequal;ity to find how many times a true coin
tnust be tossed in order that the probability will be at least 0.90 that is X,
the proportion of the number of heads will be between 0.4 and 0.6
[Hint : If n is number of toss, P (!X - n pi :5: n x O. 1) ~ 1 - 1 2
4n(·1)
1
By problem 1- 2 = 0.9 => n = 250.]
4n(-I)
51. For the discrete districution X = i
;; = P(X = i) = Iii i = 1,2,3,··· prove that Chebycheff's inequality
1
gives P (Ix - 21 > 2) < ~ while the actual probability 16.
52. A random variable X takes the values -1, 1,3, 5 with associated
2.
x a 0
3
1/8
t, :0 1
1 3
- - - -
2 3
3 1
not indep.
8 8 8 8
1 3/8 0
2 3/8 0 t.: 1 3
3. X 0 2 X: 0 2
1
25
- 10 1
-_. -
o 4/9 2/9 1/36 fx
36 36 36 ' 9
2/9 1/18 0 Y 0 1 2
J
DISCREET JOINT DlSTRlBUTION & CHEBYSHEV'S INEQUALITY 191
X2
° XI
°
3 1
0
3 1
o 9/16 3/16 IXI -
4
-
4
IX2 : -
4
-
4
3/16 1/16
5. 0 2
Y
1/8 1/8
°
1 1/8 1/4
°
1/8 -
1
not indep.
4
2 0 1/8 1/8
2 1 1 2 1
- and IYj
9 3 3 9 9
X and Yare not independent.
7. (X,Y) assumes the value (i,i) where i=0,1,2;i=0,1,2,3. The
joint probabilities are
8! 1 10! 8! 9 135 7
18'72'(1 19. 48x61O 20. 72x68 21.5022'204823'2
192 ENGINEERING MATHEMATICS -11A
24. v. 0 1 2 3 X2
~l 0 1 2 3
3 1/ 27 0 0 0
Cov(N,Xl) = 0; Cov(XpX2) = -1/3
25. (i) Y 0 1 (ii) X :0 Y :0
1 1 1 1
o 1/4 1/4 PXi: 2"2 PYi: 2" 2
1 1/4 1/4
(27
28. (i) 2 (ii) 2 (ill) 4 (iv) 0 (v) 1.2.33. Vn
3 28 45
35. "4 37. 253 38. 54
15 3
40'17 43. "4 44. no /
3
48. ·0951 49. (i) "4 (ii) 10 50.250
(c) 0
Var(A~) +X2)=
1 19 9
(a) 20 (b) 20 (c) 20 (d) none of these
14. For a random variable with mean m and s.d. a p(IX - ml ~ 2a) $;
1 1 1 1
(a) - (b) - (c) - (d)-
4 2 3 5
15. For a random variable with mean 0, p( -40- < X < 40-) ~
1 1 15 13
(a) 16 (b) -
4 (c) 16 (d) 16
16. For a random variable with mean m and variance 2,
P(lX -ml$;2)~
(a) .9 (b) .3 (c) .2 (d) ·5
17. If XI' X 2'·· .,X; be n independent variates having same distribution
with common mean 4 and variance ·01 then the destribution of
- 1 n
X =- L Xi has distribution with approximate variance
n i=1
6.1. Introduction.
In a previous chapter we defined continuous random variable and its
distribution function. Useful properties of distribution function were also
discussed in that chapter. In fact probability distribution of a continuous
random variable is given by its Probability Density Function which is
being discussed in this chapter. Concept of Expectation, Mean, Variance of
continuous random variable are nothing but the consequences of the
appearance of this density function.
6.2. Prebabillty Density Function (or density function)/
.. -
.and Continuous Distribution.
Let X be a continuous random variable with distribution function
F(x) = P(-oo < X $ x) Then a function f(x) is said to be probability
d~nsity function (pdj) of X if f( x) is integrable on the interval [a, x] for
all a and if
x
F(x):: Jf(t)dt
This holds for every real x.
Jf(t)dt
00
Properties of pdf.
(i) As F( x) is a continuous function, so we must have
P(X=a)=F(a)-limF(x)=O.
x-+a-
(ii) For a continuous distribution,
P( a ~ X ~ b) = P( a < X < b) = P( a < X ~ b)
b
= P(a ~ X < b) = f(t)dt f = F(b) - F(a),
a
00 2 P2 ()
=f-3dx= Lt f-3dx= Lt 1---;- =1-0=1
x P-+oo X P-+oo P ,
1 1
L
CONTINUOUS PROBABILITY DISTRIBUTION 197
E(X) = fxf(x)dx,
--«l
provided the infinite integral converges absolutely.
Similarly, the mean of a function ",(X) of the random variable X
denoted by E{ ",(X)} is defined as
co
Illustration.
Suppose X and Y be two continuous random variables. Their distribution
is given by the density function.
f(x) = 2x, 0<x ~ I
= 4x - 2x, I<x ~ 2
=0 , elsewhere
I
and </l(y) = -, - 2~y ~ 2
4
=0 ,elsewhere
o 1 2 OCI
1 2 2
= 2 f X2dx + 4 f xdx - 2 f X2dx =2
o 1 1
OCI 2 ~ I OCI
=![y2]2 =.!.(4-4)=O
4 2 -2 8
Then E(2X - Y + 3) = E(2X) - E(Y) + E(3)
= 2E(X)-O+3 =2x 2 = 4
We define variance and standard deviation for a continuous random
variable as that for discrete random variable given in an earlier chapter.
The positive square root of Var (X) is called the standard deviation
of X and is denoted by cr(X) or c x or simply cr. Thus c = +~var(X) .
Remarks: (i) The variance describes how widely the probability masses
are spread about the mean i.e it gives an inverse measure of concentration
of the probability masses about the mean which is called the measure of
dispersion.
(ii) As Var(X) = 0 only when X - m = 0 i.e, X = m , so in that
case the whole mass is concentrated at the mean.
Theorem.
2
(i) Var(X) = E(X2) - m2 = E(X2) _ {E(X)}
(ii) Var(aX + b) = a2Var(X)
(iii) Var(k) = 0 where k is constant.
(iv) Var(X)= E{X(X-l)}-m(m-l) where m is mean of X
Proof: (i) Var(X) = E{(X - m)2} = E(X2 - 2mX + m2)
= E(X2)- E(2mX) + E(m2)
=E(X2)-2mE(X)+m2 =E(X2)-2m.m+m2 =E(X2)_m2
(ii) and (iii) are left as exercise.
:(:;:
f~:;:
:fj:- 0
l~ x dx ~ [X:]2
0
4
3'
200 ENGINEERING MATHEMATICS-llA
Now, J
-00
J
E(X2) == x2f(x)dx == X2 '~X dx ==[X;
0
]2 ==2
0
1
:. Var(X)==E(X2)_{E(X)}2 ==2_ : ==!.
2 2
2 3
. [3X x
i.e., k T-s-2x
]2
1 =1
k
i.e
.. , -=1
6 ... k-6
- .
x x
J
Also F(x) = f(x) dx + f(x)dx,
-00
J
1
for 1 s x ~ 2
x x
= J O·dx+ J 6(x-l)(2-x)dx
-00 1
1 2 x
Again F(x)";'
-00
J O·dx+ J 1
6 (x-l)(2-x)dx+ J O·dx,
2
for x > 2
3X2 X3 ]2
= 0+ 6 [- 2 --
3
- 2x + 0 = 1
.
1
F( x) = 0 , for x <1
p( : ~ X s %) = F( %) - F( :).-
={5-12%+9H)' -2 (%)'}-{5-12 : +9(:), -2(H)
1 5 11
=---=
2 32 32
Alternative by (Using p.d.f) :
3 3
a
f(X)=-2-' -oo<x<oo·
. x +1
1
Find (i) a (ii) the probability that X2 lies between 3" and 1 (iii) the
distribution function of X.
00
+X2 dx 1
=1 -<Xl
-00
1-~ dx 11 dx
=; f 1+x2+;f
-1
1+x2
1
~
1
= -[tan
1t
-1 x]-~
~
-1
1
+ -[tan
1t
x]- -1
1
1
= ~(-~+:)+~(:-~) = ~.
(iii) The distribution function of X is given by
F(x)=- 1
1t
fX --=-
dx
1+ X2
1[tan
1t
_l]X x
-<Xl
1(tan-
=-
1t
1
x+- 1t)
2
.
-<Xl
1
f() x =-e - :0 ' x > 0
40
CONTINUOUS PROBABILITY DISTRIBUTION 203
Find (i) P(X ~ 20) (ii) P(32 ~ X s 48) (iii) P(X ~ 25)
(i)
wI
P(X s 20) = [ 40
_~
e 40dx
[_~]W = ~
= -e 40 0 1- e 2
X ]48 4 6
(ii) P(32 ~ X s 48) = 1]...e -:Odx = [
-e - 40 = e-5 - e-5
~40 ~
25 1 _~
(iii) P(X ~ 25) = I-P(X < 25) = 1-
o
J
40 e 40dx
1. -2 1
.-2 1 1 1 3
= JOdx+J-dx=-(1+2)=-
--<lO -2
444
1
2
J J
00
= f(x)dx + f(x)dx
3 -00
21 21
= J-dx+O+O+ J-dx
4
34 _2
=~(2-%)+~(~+2)=!'
Example 5. Consider the p.d.f f(x) = ae-b\x\ where x is a random variable
whose allowable range of value are from x = -00 to 00. Find (i)
cumulative distribution function (ii) the relation between a and b (iii)
P~sxs~. .
(i) The cumulative distribution function is given by
x x
F(x) = J f(x)dx J
= aebxdx, for -00 <X <0
-00 -00
o x
o
= J aebxdx + J ae-bxdx
x
a [bX)O
=- e
b -00
a [ _bX]X
+--e
b 0
-00 0
CONTINUOUS PROBABILITY DISTRIBUTION 205
a
(ii) Since F(oo) =1, so, -,;(2-0)=1
. 2a = 1 . _
i.e., b .. b - 2a
(iii) Now P(l s X s 2) = F(2) - F(I)
-b
= ~ (2 - e -2b - 2+e -b) = e2 (1- e -b) .
Example 6. Show that a function which is Ixl In (-1,1) and zero
elsewhere is a possible p.d.f and find the corresponding distribution
function. [W.B.U.T. 2013,2006]
Let us denote the given function by f(x). Then
f(x)=lxl in -1<x<1
= 0, elsewhere
i.e., f(x)=-x, -1<x~O
=x, O<x<1
= 0, elsewhere
Clearly we see that j(x) ~ 0 everywhere.
00 -1 0 1 00
-1 x
F(x) = f f(x)dx + f f(x)dx, for -1 < x s0
~ -1
-1' x
= fo.dx+ f(-x)dx=i(I-X2)
-'>' -1
206 ENGINEERING MATHEMATICS - rIA
-lOx
F(x) = I f(x)dx + I f(x)dx + I f(x)dx,
-<X> -1 0
for °< x :5 1
. 0 x 1
=0+ I(-x)dx+Ixdx="2(1+x2),for 0<x:51.
-1 0
-1 0 1 x
o 1
F(x) = 0, -oo<x:5-1
~ ~(I-X2), -1<x:50
1 .
=2"(I+X2), 0<x:51
=1 l<x<oo
Example 7. X is a continuous random variable having p.d.f
41 22 17
= -fx2dx +-fx(3-x)dx =-
5 5 15
o 1
CONTINUOUS PROBABILITY DISTRIBUTION 207
= fX2{1-(1-x)}dx+ fx2{1+(1-x)}dx
o 1
00 0 2
E(X3) = f x3f(x)dx 3
= f x ·0 dx + f x3{1-ll- xl}dx
-<X) -<X) 0
00 2
+f x3 -txix = f x3{1-11-xl}dx
2 0
208 ENGINEERING MATHEMATICS-IIA
. 3 1.2 3
1
Mean = E(X) = f x f(x)dx = f x·~dx+ f x·-dx +fx.! (3 - x)dx
. 2 2 2
. 0 0 1 2
1 1
=-(1-0)+-(4-1)+-
6 4 2 2
(27
1 --9-6+- 8)
3
=-
3
2
3 123
8
=-
3
= 0, x <0
Find the mean, variance of this random variable and also find the probability
that it will exceed A.
J
00 X2
X2
J X2,
00
= 2~2J
·002
I'v U
2-1
e-ud u, by puttmg
,x
2/..2 = u
o
= 2,,2r(2) = 2,,2 [ ',' r(2) = I! ]
X2
= 1- Je
2 -u
_1:.J -1:. 1
=1-1-e2 ( =e2=J;'
I
--<X) 0
J
00
J (x _1)2 e-xdx
00
2 x
= Jx e- dx-2Jxe-xdx+Je-X =1(3)-2 1(2)+r(1)
o 0 0
= 2!- 2 . 1+ 1= 1
U
= Lt
U--+oo
Je -3d:X;
x
= Lt (_3e-U/3 +3) = 0 +3 = 3.
U--+oo
o
Example 12. The distribution function of a random variable X is
F(X) = CX3, O~x<3
= 1, x~3
= 0, x<O
If P(X = 3) = 0, find (i) the constant c (ii) the density function
(iii) P(X > 1) (iv) P(l < X s 2) (v) P(3X + 2 < 8)
(i) The density function/(x) is given by f(x) = F'(x)
.. f(x) = 3CX2, 0~x < 3
= 0, elsewhere
J f(x)dx
00
Also we have = 1
-00
i.e., 27c =1 1
... c=-·
27
= °, elsewhere
CONTINUOUS PROBABILITY DISTRIBUTION 211
00 . 1 3 26
2
(iii) P(X > 1) = ff(x) dx = -fx dx =-
9 27
1 1
2 1 2 2
2
(iv) P(l < X ~ 2) = ff(x) dx = -fx dx =-
91 27
1
4
1[X5 3x 3 X2J3
="4 5-7+x -2 =2.65.
1
is a possible density function and compute p( X > ~). Also find E(X)
" 00
ie., kf x(l-x)dx =1
o
X2
i.e., k [ 2-"3
x3l1 =1
0
i.e., k(i-~) =1
:. k=6.
.. f(x) = 6x(1-x),0 < x $1
=0, else where.
Now, E(x)=6fX2(I-x)dx
o
<J:
~{~3
=6(j- ~)=i·
Exercise
[I] Short Answer Questions
1. Find mean and S.D of the following distribution:
1 -~ .
f(x) = -e 4,X> 0
4
=0 , else where.
.. .. 1~i ~
x O<x~1
2. Let f(x) = 2- x 1sx ~2
{
. 0 elsewhere
0 ,x < 2
f(x) = ~(3+2x),2 ~x ~4
{ 18
o ,x> 4
a density function ?
214 ENGINEERING MATHEMATICS-ItA
= 1].
~ 1 -=-
5. Verify that the function f(x) = -e
')...
A , o<x<oo
= 0,
elsewhere
is a possible probability density function. Find P(2 < X < 6) .
6. (a) Ifa random variable X has the density function
1
f(x) = "6' -3 < x < 3
=0, elsewhere,
obtain (i) P(X < 1) , (ii) p( IXI > 1),. (iii) P(2X + 3;> 5)
(b) A random varaible X has the following pdf
f(x)=k, -2<x<2
= 0, otherwise.
(i) Determine the constant k (ii) what is the value of P(lxl > 1) ?
[W.B.U.Tech 2004]
(c) Obtain E(X), Var(X) of X whose pdf is given by
f(x)=2x,O~x~1.
= 0, otherwise.
Hence find Var( 3 - 5X) .
7. Is the following a p.d.f?
f(x)=2x, O<x~1
= 4-2x, 1<x~2
elsewhere.
8. If the random variable X has the p.d.f
1 .
f(x)=4·-:-2~x~2
2 2 1 2 1 25
6. (a) (i) 3" (ii) 3" (iii)"3 (c) 3' 18'
18 7. not p.d.f.
1
9. k = 6, mean = 9.5
10 .! J8
8. 4' . 3' 3 '
1
f(x)=2' 4~x~6
= 0, elsewhere.
Find (i) P( 4 ~ X s 5) (ii) P(X < 3.2) (iii) P(X < 4.2)
(iv) P(3 s X s 4.5) (v) P(X > 5.5) (vi) c.d.f (cumulative distribution
function i.e., distribution function) (vii) k such that P(X < k) = 0.7
5. The p.d.f of a random variable X is
f (x) = ex 2, 0sx s1
=0, otherwise. ...;.-
Find (i) e (ii) ~O~X~~) (iii)P(4X > 3) (iv) P(l < 4X < 3)
6. f(x) =~ -
kx, 0 $ X $ 4 is the p.d.f of a r.v X Determine (i)
2
k (ii) the probability that X assumes the value between 2 and 3 (ill) the c.d.f (iv)
P{X$l) (v) l{X~25) (vi) *-~<Q5)
7. The length of the life of a tyre manufactured by a company
follow a continuous distribution with density function
k .
f(x) = -3,1000 $ x $150.0 .
x
Find k, the probability that a tyre selected at random would function
for at least 1200 hours.
8. The p.d.f of a r.v is f(x) = k(x -1)(2 - x), 1 $ x $ 2. Find (i) k
5
(ii) c.d.f (iii) the probability that X is less than '4 (iv) the probability that
L
218 ENGINEERING MATHEMATlCS- llA
O<x~l
= 2' 1~x~2
1
= "2(3 - X), 2~x ~ 3
(i) Verify that the area under the curve is unity
(ii) Find mean and variance of X
13. If the pdf of a r.v X is given by
-(x2+2x+3)
f (x) = ce ,- 00 < x < 00
find the value of the constant c, the expectation and variance of the
distribution.
14. (a) The demand for a new product of company is assumed to be
a random variable with p.d.f.
= 0, x < 0
Find the mean and variance of this random variable and find the
probability that it will exceed 1.
(b) The length of the life of tyre manufactured by a company follows
a continuous distribution given by the density function
k
j(x) = -3' 1000 ~ x s 1500
x
=0 , elsewhere
Find k and find the probability that a randomly selected tyre would
function for at leasfl200 hours. [W.B. U. Tech 2005]
15. A continuous distribution is given by
f(x)=_I_e-(logx)2/2 x>0
x..{i; ,
= 0, x <0
Find the mean and standard deviation of the distribution.
CONTINUOUS PROBABILITY DISTRmUTION 219
oo 2 J;.
[Assume
o
f e-X dx = 2]
16.The pdf of a continuous r.v is given by
f(x) = Ke-b(x~a) ,a ~ x < 00
= 0, elsewhere
. 1
wha"e K, a, b are constants. Prove that K = b = - and a
. (J
=•.in - (J
.
.
[Hint: Var(X) = - - - k2 ]4(~)
9 4
~swers
3e-4 e-2 1 17 5
1.1'--3-'-- 2. 25 '25'
e e
x-4
(v) 0.25 (vi) -2- (vii) 5.4
1 3
S. (i) 3, (ii) 8' (iii) 37/64, (iv) 13/32, (v) x
.. 3 ... x(S - x) . 7 9
1
1 (vi) -
6. (i)8 (ii) 16 (iii) 16 (iv) 16 (v) 64 4
7. 36x105 ; ~.
20
5 1 (v) .!.!..
8. (i) 6 (ii) 5 + 9x2 -12x - 2x3 (iii) --
32 (iv)-2
32
2 .
6x - x - 5 , 2 < x <
_ 3 .' 1, x > 3
4
71t 341C2
11.3'45
14. (a)
r; 4-1t
V"2' -2-'..{e
1
(b) 0.45
IS . ..{e, ~e2 -e
18.0
CONTINUOUS PROBABILITY DISTRIBUTION 221
f(x) = e-x , x ~ 0
= 0, otherwise.
Then the value of E( X) is
(a) 0 (b) 1
(c) 2 (d) -1 [W.E. U. Tech.2007]
222
ENGINEERING MATHEMATICS -IIA
= 3- 2
X , for 2:$ x :$ 5
=1, for z » 5.
Then P( 3 < X :$ 4) =?
(a) -7 (b) 7
(c) -13 (d) 6
7. A random variable X has the following p.d.f:
f(x)=k,-2<x<2
= 0, otherwise.
Then the value of the constant k is
1
(a) - 1
8 (b) -
2
1
(c) - 1
4 (d) 12 [W.B.U. Tech. 2007]
= A,5 :$ x < 00 .
(c)-_
2 1
3 (d) 3"'
9. Let X be a continuous random variable with distribution
f()x = .!.8'
{
° :$ x s8
0, elsewhere
CONTINUOUS PROBABILITY DISTRIBUTION 223
Then P(X s 6) =
1
.l 3
(a) - (b) -
4 4
7 1
(c)-
8 (d) "2'
10. The following distribution is a p.df of a random variable X :
(c) ° (d)
1
3"'
12. The distribution of a random variable X is
F(x) = 0, x<O
= X2 + 3, 0 ~ x < 1
=1, x~l·
Then the corresponding p.d.f of x is
x3
(a) - + 3x, 0 ~ x < 1
3
0, else where.
0, else where.
224 ENGINEERING MATHEMATICS - IIA
13. Which one of the followings is not true for a distribution function
(b) lim
x-t-ao
~(x) = 0
(a) 3crx
(c) 3crx + 10
CONTINUOUS PROBABILITY DISTRIBUTION 225
(a) 6 (b) 5
(c) 36 (d)-6
EM-2A-IS
226 ENGINEERING MATHEMATICS-lIA
24. If P(X = i) = p and P(Y = j) = q,(i,j = 1,2) and X and Yare two
mutually independent random variables, E(XY) =
1 7
(a) - (b) -
4 4
9
(c) - (d) none of these.
4
25. A random variable X has the pdf
1
f(x)=-,-2<x<2
2
= 0, otherwise.
1 1
(a) - (b) -
3 2
1 1
(c) -
7
(d) '9'
Answers
l.b 2.a 3.b 4.b 5.b 6.a 7.c 8.a 9.b
7.1. Introduction.
In an earlier chapter we introduced some special type of discrete
distribution. In this chapter similarly we are giong to introduce some special
type of continuous distribution. Three such type of distribution will be
discussed in this chapter.
7.2. Normal Distribution.
A continuous random variable X is said to have a normal distribution
if its probability density function is given by
(X_I!)2 .
1--2-
.. f(x)=--e
cr..J2; 20' ,-oo<x<oo·
-<Xl
00
-<Xl
2(72
dx
1
= r Je
00
-z
2 •
dz , by putting z = x-IlIn
v7t -<Xl
crv2
J
00
1 e -u u 2d u , by putting z 2 = u
= 2I e- z2 dz = r J
00 1
v7to v7to
= ]; r(~) = ];. ~ = 1.
So the two necessary conditions for the probability density function
are satisfied.
(3) The significance of the parameters Il and c are given in the next
theorem.
228 ENGINEERING MATHEM~TICS-IIA
--«J
=-
crJ2;.
J X
--«J
e 20
2
--~--~--~--~~----~----~~--~---------+x
The normal curve is bell shaped and symmetric about the ordinate
x = Il . For small values of o , the curve has a small peak and as o
increases, the normal curve tends to be flatter.
Theorem 1. If X has normal distribution with parameter Il and o then
(i) the mean of X is' Il (ii) the s.d of X is cr.
Proof Here p.d.f of X is given by
(x_I')2
1 --- 2
f(x) = r;:;-e 20 , -oo<x<oo
crv21t
00 (X_I')2
(i) Mean = E(X)=_l_Jxe-2T dx
crJ2;. --«J
(x_I')2 (X_I')2
= _1_
crJ2;.
J (x - -2T dx + crJ2;. J -2T dx
00
Il)e _Il_
00
e
--«J --«J
= r J ue - d u + r J - d u
00 00
u2 u2
1 Il x - Il
u = -- e. by putting
1tv --«J 1t v --«J' cr~
SPECIAL TYPE OF CONTINUOUS DISTRIBUTION 229
J
00
2
= 0 + ~ 2 . e _u du [ '.' the integrand of the 1st integral is odd
'lire 0
and that of the 2nd integral is even]
=~ J Z2- e-zdz,
f;.o
JC 1
1
by putting u2 =z
= j;r(~)= j;f;. = ~
(ii) Now, Var(X) = E{(X _ ~)2}
= f;. J z
2 c2 00 x-
2
e
-z2d
z by puttmg
.
crF2
~
=z
--«>
2
cr
= f;.' f;. = cr 2
~~----------~o~-----------=~-
x
It is symmetric about y axis, bell shaped. It has maximum value at
asymptote.
Standard Normal Distribution Curve.
The graph of the distribution function ~(x) is shown in the following
figure.
<t>(x)
____------~----------------------~x
o
SPECIAL TYPE OF CONTINUOUS DISTRIBUTION 231
Z=E(Z)=E(X~I-l)= ~{E(X)-E(I-l)}
=l.-{I-l-I-l}=O
o
=~{(j2_0}=1.
c
Tabulation ofthe Standard Normal Distribution.
Let the random variable X has standard normal distribution. Then its
distribution function
x
<t>(x) = P(-oo < X s x) =
-<Xl
J <j>(z)dz
=-
1
x
.s:
2
Je 2dz:. F{X~a)=-
IQZ-
,
An Illustrative Examples
Let a thermal plant supplies electric power in a certain city. Let
X = amount of electric power (in watt) supplied by the plant in a day.
Clearly X varies from day to day.
It can be assumed that this X has normal distribution with parameter
~= average power supply per day and o = standard deviation of all the
values assumed by X Now ifit is known that these m = 300 and c = 10
. X -300
then by the preVIOUStheorem Z = is a standard normal variate.
10
Suppose we are asked to find the number of days on which power
supplied will lie 'between 280 to 310 MW.
Then we find P(280 < X < 310)
= f
I
t. = (n).
i pt (1- p) n - i , i = 0,1,2,· ..,n
fO=12CO(0.5)o(0.5)12-0 = 0.000244
.20
.15
.10
.05
o I 2 3 4 5 6 7 8 9 10 II 12
We think n as large. So X is approximately normal variate with
parameter J..l = np = 12x.5= 6 and
l.e., Z =
X - 6 is approximately a standard normal variate. In the
1.73
above figure the curve is (6, 1.73) normal curve. This approximately fits
= J <j>(t)dt + J <j>(t)dt
-1.46 0
= 0.4279 + 0.3078
(obtained from the statistical table-I)
= 0.7357·
(ii) Now, suppose we seek P(X = 8).
We calculate this with the help of approximate normal distribution then
we find P(7.5 ~ X ~ 8.5) = P(0.87 ~ Z s 1.44) which would be
.. P(X>c)=P
X -3
( -2->-2-
C-3) =P Z>-2-
( C-3)
c-3
c-3
2
X -3 c-3 c-3
Also P(X 5c) = P( -2- 5-2- ) = P( Z 5-2- ) = J <I>(t)dt
-00
-00
c-3
or, -2- =.43 (from the given data)
:. c = 3.86.
Example 3. The length of bolts produced by a machine is normally
distributed with mean 4 and and s.d. 0.5. A bolt is defective if its length
does not lie in the interval (3.8, 4.3). Find the percentage of defective bolts
produced by the machine.
1
Ie'
2
\ 1
lJ2;; y'
0.6 e eft
~O.7257, J2;;
1 0.4 t eft
~o.6554j
[WE. U.Tech 2004]
= f <I>(t)dt
-0.4
-0.4 0.6
0.6 -0.4
f f <I>(t)dt = f <I>(t)dt [.:
co .
= 0.7257 -
(
1- L
0.4 J
<I>(t)dt = 0.7257 - (1- 0.6554) = 0.3811
:. Probability that the length of the bolt lies between 3.8 and 4.3
is 0.3811.
:. Probability that the length of the bolt does not lie between 3.8 and
4.3 = 1- 0.3811 = 0.6189 .
:. Probability that the bolt is defective = 0.6189
f
z
[Given that ~ e-t2/2dt = 0.1554 and 0.2881 according as z = 0.4
-V21t 0
and z = 0.8]
SPECIAL TYPE OF CONTINUOUS DISTRIBUTION 237
-O~ OA
= area under st. normal curve enclosed between the two ordinates
Z = -0.8 and Z = 0.4 (shown by shade in the figure)
0.4 0 0.4
0.8 0.4
= f f
<j>(t)dt + <j>(t)dt, since <I> curve is symmetric about Yaxis.
o 0
0.8
= 0.5 - J <p(t)dt = 0.5 - 0.2881 = 0.2119
o
:. the expected number of workers = 0.2119 x 10,000 = 2119
(iii) When X = 72, Z = 0.4
00
From the supplied data we have P(0 < Z < 1.64) =.45
or, P( Z > 1.64) = 0.5 - 0.45 = 0.05 (2)
10
Comparing (1) and (2) we get -
CJ
= 1.64
10 :. the s.d is 6.1.
or, CJ = -- = 6.097 ~ 6.1
1.64
Example 6. In a normal distribution, 31% of the items are under 45 and
8% are above 64. Find the mean and standard deviation. [Given
P(O < Z < 1.405) = 0.42, P( -0.496 < Z < 0) = 0.19
[ W.B. U.Tech 2003 ]
SPECIAL TYPE OF CONTINUOUS DISTRmUTlON 239
From above
Then
Z = x - Il = X -12 :. X = 4Z + 12
cr 4
:. (i) P(X? 20) = P(4Z+12? 20) = P(Z? 2) = I-P(Z < 2) = 1-<D(2)
= 1- 0.9772 = 0.0228·
(ii) P(O ~ X s 12) = P(O s 4Z + 12 ~ 12) = P( -3 s Z s 0)
= P(O s Z s 3) (due to symmetry)
( a ~ 12)
or, "1- <1> = 0.24 , where <1>is c.d.f of st. normal variate
variance of X •
Let E(X) = Il, Var(X) = cr2 • Now, if <1>is c.d.f of st normal variate
<1>(x)= p( X; Il ~ x) = kJ U2
e- 2du
-00
/
1
SPECIAL TYPE OF CONTINUOUS DISTRmUTION 241
= P(9.6 ~ X ~ 13.8)
. Z = X - 200 IS
i.e. . approximate
. stan dar d norma I'variate.
10
.
i.e., Z = X -1000 .
IS .
approximate Iy stan d ar d norma 1·vanate. S·mce
30
discrete variate is approximated by continuous variate so we find
P(X ~ 950.5).
Now X = 950.5::::}Z = -1.65
1.65
1.65
1 --x-a
f(x)=be b ,x~a
= 0, elsewhere
where a, b (b > 0 andb > 0) are two parameters of the distribution. f(x) is
called Exponential Probability Density of X.
ote. (1) Clearly f( x) ~ 0 for all x and
oo Joo1 -~
. f(x)dx be dx 1.
J
-00
=
a
b =
f(x)
1
b
o a
244 ENGINEERING MATHEMATlCS- UA
F(x) = 0, -co c x c c
x-a
=l-e b ,
(0,1) --------------------------------_.
__+- -+ -+x
o a
probability density is
= 0, elsewhere
where A > 0 is the only parameter.
ote. (1) By the termX has exponential distribution' we mean X
has one-parameter exponential distribution.
(2) Clearly the pdf f(x) satisfies the two fundamental properties of pdf.
(3) If X has exponential distribution with parameter ).." we write
X - E(O,)..,,).
(4) This distribution is obtained from the previous by putting
a=O , b=! ).." .
Theorem. If a continuous random variable X has exponential distribution
with parameter ).."then
1 1
[w.B.U.r. 2013]
(i) the mean is i (ii) the variance is )..,,2 •
SPECIAL TYPE OF CONTINUOUS DISTRlliUTION 245
1 ""
=ir(2) [':r(n)= fe-Xxr1dx from definition ofGanuna function]
o
1
= '" [.: r(n + 1) = n! when n is a +ve integer]
(ii) Now, f
""
E(X2) = x2)..e-udx =
o
~Ju2e- du
A
""
0
u byputting Ax =u
=~r(3) =~ =.!.
",2 ",2 ",2
:. Var(X) = E(X2)_m2 [where m is mean]
Illustrative Example.
Example 1. Suppose that during rainy season, on a tropical island, the
length of shower has an exponential distribution with average length of
1
shower '2 mins. What is the probability that a shower will last more than
three minutes ? If a shower has already lasted for 2 minutes, what is the
probability that it will last for at least one more minute?
Let X = length of shower in minute.
1
P((X ~ 3)n(X ~ 2)) P(X ~ 3) ~ 1
== P(X ~ 2) ==P(X ~ 2) ==T ==~ .
e4
Example 2. Let the length of a phone call (in minutes) is exponentially
distributed with mean 10. If some one arrives immediately ahead of you at
a telephone booth, find the probability that you will have to wait for
(i) more than 10 minutes
I 00 I
-~ x -~ I 10 I
==- Ie l°dx=_lim Ie lodx=_._==_
10 10 10 x-+0010 10 e e
Ui) The required probability ==P(IO < X < 20)
20 I .s. I I
== I-e 10 = _
1010 e e2
I
SPEClAL TYPE OF CO TINUOUS DISTRIBUTION 247
Example 3. Let the kilometers that a car can run before its battery wears
out is exponentially distributed with a mean 10,000 km. If a person desires
to take a 5000 km trip, what is the probability that he will be able to complete
the trip without having to replace the car battery ?
Solution. Let X = the remaining life time (in-thousand kilometer) of the
battery. Therefore X has exponential distribution with parameter A = ~ .
The required probability = P(X > 5) 10
x; 1 .s.
= ff(x)dx wheref(x)=-e 10
5 10
1 co .s. 1 x .s. -.!.
=- Ie l°dx=-lim Ie l0dx=e 2
10s 10X-+oos
1
Example 4. If X has exponential distribution with mean A prove that
ECXI/) n! h . . ...
=- were n IS a positive integer.
An
co
Solution. E(Xn) = I xl/f(x)dx where f(x) = Ae-h
co 00
co
J
t"
= An e-/ dt putting, Ax = t i.e., dx = ± dt
CO
AI co tl-! 1
= -- fe-I -(-I . -dt, putting AX = t.Xdx = dt
reI) ° A- A
---- 1 _ 1 r(/)-1_
e -I t Hd t---
°f
co
r(1) reI)
So the two fundamental properties of pdf are satisfied.
(2) The gamma distribution with A = ..!. and I =!2 n is a positive integer
is given by the probability density 2 2
-- 1
e x2 ( -x )~-I7
-
2
f(x) = () x~o
zr !2
2
=0 , elsewhere
=0 , elsewhere
which is nothing but the density function of one Parameter Exponential
distribution.
=- 1 ""J -z"1 1 d
I'(I) 0 e z -A z'" [Puttmg AX= z .,. Adx = dz ]
1 1
=-Zr(1) =-
Ar(1) A
eo
1 ""J -z 1+1 dz
= Ar(1) 0 e z ;: Putting AX = z :. Adx = dz
1
=-r(/+2)
A2r(1)
250
ENGINEERING MATHEMATlCS-UA
1
= A2r(/) (l + 1)1r(l) [by a property of garrnna function]
1(/ + 1)
=)}
= l(l + 1) _
,1.2
(i)2 =
A
l(l + 1) _ ~
,1.2 ,1.2
,
/1.2
.. Standard deviation of X =~ .
Gamma denisty Curve
The graph of the pdf of a gamma variate is shown in the following
figure for different values of' and A
Y f(x)
1 ,= 1, ,1.=1
'.
".
--O~--------~4~=-~6~~~·~~~~·--------~X
Cases where Gamma distribution fits
Let there be n number of events occuring randomly. X = the time at
which the nth event occurs. Then it can be shown that the variable X has
gamma distribution with parameters ., = n and A ·where A is average
number of event occuring in unit time interval.
That is the pdf of X will be
Ae-..lX(Axr-1
f(x) = (n-I)! ,x>O
SPECIAL TYPE OF CONTINUOUS D1STRmUTION 251
I =2 and A =.!..,
find the mean and the standard deviation of this
3
distribution. Hence find the probability that the random variable will take
a value less than 5.
o I 2
Solution. The mean, E(X) = - = -1 = 6
A .
The standard deviation = Ji = T2 = 3J2
A 1/3
The pdf of the distribution is
1 -~
-e 3 -
(X)2-1
3 3 x;::: 0
f(x) = 1(2) ,
=0 , elsewhere
=0 , elsewhere
5
So, the required probability = P(X < 5) = J f(x)dx
=
5I~xe
1 --x
3dx~"9
1 {[ --x]5
-x·3e 3 0 -
5 --x}
1-3e +dx
=i{-lse-l-3+ ;I}
="91 { -15e -~3 -ge -~3 +9 } =-ge
24-~
3 +1=0·496
252 ENGINEERING MATHEMATlCS-IIA
I 1
:.12A.2 =6A. :.A.="2. So,I=6x"2=3
. 1 _.!.x (x J-1
:. thep<!fofX is f(x)~ 2' ~(3~ , xz O
=0 elsewhere
.
i.e., 1
f() x =-x 2-~
e , x~o
16
=0 elsewhere
f f(x)dx
<0
=~ lim
16 x--+«> 12
'f x2e-~dx =~ lim
16x-+<o
{-2[x2e-~lX + 41xe-~dx}
. 12 12
1 400
= 16 x7 (detail evaluation is not shown)
= 0·062
:. No. of days the town will suffer load shedding
= 0 . 062 x 3654 = 22 . 63 = 23 days.
SPECIAL TYPE OF CONTINUOUS DISTRIBUTION 253
Example 3. Show that when I> 1, the graph of gamma density has a
local maximum at x = I -1 .
A
Ae-A.X(Axi-1
Solution. The pdf is f(x) = T(I) x>0
or , f() X
AI
=-x H-A.x
e
I'(I)
= ~xl-2e-A.X(-Ax+l-l)
reI)
I-I
:. f'(X) = O=> -Ax+/-l =0 => x =-
A
/-1
.: x >O,e> 0 so when x <-,f'ex) >0
A
/-1
and when x> - ,f'(X) < 0
A
. . . /-1
:. f( x ) IS maxunum at x = - .
A
Exercise
= 0.4238]
254 ENGINEERING MATHEMATICS - UA
:: xlim[e-X]X
-1
->00
:: -lim(e-X -1) = -lim(~-I)
X
x->oo eX ->00
o
;:;-(0-1):: 1]
ANSWERS
1
1.0.4238 2.0.921 3.(i) .7654 (ii) .3174 4. ",2 5. 1
1 2
[Given J <j>(t)dt =.0398, J <j>(t)dt = 0.4772]
o 0
2. A sample of 100 dry battery cells tested to find the length of life
produced the following results : x
= 12 hours, (J = 3 hours. Assuming
that the data are normally distributed, what percentage of battery cells to
have life (i) more than 15 hours (ii) less than 6 hours (iii) between 10 and
z
14 hours? Given J <j>(t)dt = 0.4938,0.4772,0.3413 and 0.2487
o
according as z = 2.5,2,1 and 0.67.
1 2 _~ 1 0.4 _~
Given ~
v2n
Je
~
2 dt = 0.9772, ~
v2n
Je
-00
2 dt = 0.6554
'
2.4 t2
kJ
1 x2
6. The mean and s.d of the I.Q of a group of 500 children is 90 and
20 respectively. Assuming that the I.Q is normally distributed, find the
J2;. f--
00 X2
1
number of children with I.Q ~ 100. Given that e 2 dx = 0.308
1t 0.5
<1>( z) =
1
.J2; o[ Z2 J
exp - 2 between z = 0 and z = 2 is 0.4772 and between
11. The mean of the inner diameters (in inch) of a sample of 200 tubes
produced by a machine is 0.502 and the s.d is 0.005. The purpose for
which these tubes are intended allows a maximum tolerance in the diameter
of 0.496 to 0.508 (i.e. otherwise the tubes are considered defective). What
percentage of the tubes produced by the machine is defective, if the
diameters are found to be normally distributed? (Area under the standard
normal curve between z = 0 and z = 1.2 is 0.3849)
12. If skulls are classified as A, B, C according as the length-breadth
index is under 75, between 75 and 80, over 80, find approximately
(assuming that the distribution is normal) the mean and standard deviation
of a series in which A are 58%, Bare 38% and Care 4%, being given
that if
It J exp (1-"2
f(t) = .[2; 1 I":
,1
X2
o
then f(0.20) = 0.08 and f(1.75) = 0.46.
17. (a) A fair coin is tossed 10 times. Find the probability of obtaining
between 4 and 7 heads inclusive by using (i) the binomial distribution (ii)
the normal approximation. (See stat. table for relevant data)
18. Among 625 random digits, find the probability that the digit 7
appears (i) between 50 and 60 times inclusive (ii) between 60 and 70 times
inclusive. (See stat. table for relevant data)
19. A pair of dice is rolled 900 times and X denote the number of
times a total of 9 occurs. Find P(SO ~ X ~ 120)
20. The ideal size of a class of a college is 150 students. The college,
experienced from past, knows that only 30% of the admitted students will
actually attend. The college uses a policy of approving the applications of
450 students. Find the probability that more than 150 students attend the
class. Area under the standard normal curve enclosed between the
ordinates z = 0 and z = L59 is 0.441.
21. One thousand independent rolls of a fair die is made. Find the
approximate probability that the number 6 appears between 150 and 200
times inclusively. It number 6 appears exactly 200 times, find the probability
that number 5 appears less thau 150 times. Use appropriate statistical table.
SPECIAL TYPE OF CONTINUOUS DISTRlBUTION 259
e 2 = 0.604)
1
exceeds 9 hours ?
26. The number of years a computer works smoothly is exponentially
distributed with vairance 64. If Prof. Das buys a used computer, what is
the probability that it will be working after an additional 8 years ?
27. Show that the graph of the gamma function with parameter I and
A has no maximum when 0</ < 1, has maxirmiin at x = 0 when 1=1.
260
ENGINEERING MATHEMATICS -IIA
Answers
1. (a) 8,200 (b) 200 (c) 200
14. 125 5
15. 2~3 16. 36
18. 0.3518, 0.5131 19. 0.96 20. .0559 21. 0.9258, 0.1762
(c) -3 1
(d) 3"'
L
SPECIAL TYPE OF CONTINUOUS DISTRIBUTION 261
1 1
(a) A (b) - (d) A2
A
3. If X is normally distributed with zero mean and unit variance, then
the expectation of X2 is
(a)1 (b) 2 (c) 8 (d)20.
[W.B.U.Tech. 2007]
6. If F( x) = 0, - 00< x <4
x-4
=--,4~x<6
2
=1>6~x<00
be the distribution function of a uniform random variable then value of
its pdf on [4,6] lS
111 1
(a) -
3
(b) -
6
(c)-
4 (d) 2'
7. If f(x) is the pdf of a two-parameter-exponential variate with
parameter 2 and 6 then, for x ~2,f(x) =
x-2 x-2
1 --x-2 1 -- (d) ~e--2
(a) -e 6 (c) -e 6
2 2 4
262
ENGINEERING MATHEMATlCS_IJA
1
(a) O~f(x)~- (b) O~f(x)~.!.
5
3
(c) O~f(X)~3
(d) O~f(x)~5.
x-6
x-6
(a) e 8 ,6~x<00 (b) 1- e 8, 6 ~ x < 00
x-6 x-8
(c) l-e-8,8~x<00 (d) l-e-6 ,6~x<00.
10. If X has exponential distribution with parameter 5"1 then its standard
deviation is
1 1
(a) 25 (b) -
5 (c) 25 (d) 5.
20. If JO'{ is a normal variate with mean 4 and s.d. 0.5 and z be standard
normal var iate then z = 0.6 implies X =
(b) 5 (c) 1.6 (d) 4.3
264
ENGINEERING MATHEMATICS -IIA
_x2(1-x 2)2
e
(b) 2
4
(e) e-f(~x2
4
r (d) none of these.
Ans'wcrs
La 2.b 3.a 4.b 5.b 6.d 7.c 8.a
9.b 10. d ll.c 12.d 13.b 14.c 15.d 16.c
17.d 18.a 19.c 20.d 21.lJ 22.d 23.d 24.c
MODULE-3
BIVARIATE DISTRIBUTIONS
m~======(;;;C;;;;;O;;;;;N;;;;;T;;;;;INU=O;;;;;U;;;;;S;;;;;V;;;;;A;;;;;R;;;;;I;;;;;A;;;;;TE;;;;;S;;;;;;)
8.1. Introduction.
In an earlier chapter we introduced joint distribution or bivariate
distribution for discrete random variables. In this chapter we present the
concept of bivariate distribution or joint distribution for a pair of continuous
random variable.
Just as discrete case if the continuous random variable X and Yassumes
values x and y in the interval a ~ x ~ b and c ~ y ~ d respectively then
Bivariate (X, Y) assumes all the values (x, y) lying within the rectangle
having vertices (a, c), (b, c), (b, d) and (a, d)
a b
°
00 00
I I-x
= f f 6(1- x - y) dy dx
x=o y=o
I ( 2 )I-X
=6 f
y-xy-~
x=o . y=o
It
=6 (1-X)-X(1-x)-
(1_X)2t~
2 r
Jl
= -6 If(2 - 2x - 2x + 2x 2 - 1- x 2 + 2x ) dx
20
=3[(X -2X+I) 2
=3[X; ->'+x1 =3(t-1+1)=1
So the function obeys the two fundamental requirement of pdf
Therefore this function is a pdf of some bivariate (X, Y).
8.5. Marginal Distribution
y 00
fx(x) = f f(x,y)dy
00
o
[1
= 6~xf (1- x - y) dy = 6 - - x + -X2) = 3(1- 2x + X2)
2 2
f (/ ) f(x,yo)
XlY x Yo = fy(yo)
.• (Ix) - f(xo,Y)
NIX y 0 -. fx(xo)
I
BIVARIATE DISTRIBUTIONS (CONTINUOUS VARIATE) 269
j, (x/O.S)= f(x,O·S)
);/y fy (0· S)
f(x,O·S)
= using definition of marginal pdf
00
J f(x,O·S)dx
-00
12
-x(2-x-0·S) x(2-x-0·S)
= S = 2 O·S
112 ---
J-(2-x-0.S)dx 3 2
o S
= 6x(2-x-0·S) = 6x(1·S-x)
4-3xO·S 2·S
The conditional density of X given Y =y will be
f(x,y) 6x(2 - x - y)
fx/y(x/y)= fy(y) = 4-3y (evaluation is as above).
Y=yo,P(XEA/Y=yo)= Jfx/y(x/yo)dx
A
7
= 0'I 6x(2-x-0· 2)
dx
0·5
4-3xO·2
300.7 30
=- I x(I·8-x)dx =-xO·143333
17 0.5 17
8.7. Independent Random Variables
In a previous chapter we defined the independence of two discrete
random variables. Going to the same line we define •
"Definition : Two continuous random variables X and Yare said to be
independent if f(x,y) = fx(x)fy(Y) V X,Y
where f(x,y) is the joint pdf and fx,fy are marginal pdf of X, Y
respectively. "
Theorem 1. If X, Yare independent random variables and A, Bare
two events then b d
Pea < X ~ b, C < Y ~ d) = Ifx(x)dx Ify(y)dy
a C
fu v(u, v) = fx Y(X,y)18(X;y)1
' . 8(u,v)
Proof. Beyond the scope of the book.
This theorem leads us to obtain the distribution of sums and quotients
of two random variables as discussed below:
8.8. Distribution of Sums and Quotients of random variables
Theorem 1. If X and Yare independent continuous variates then the
probability density function of U = X + Y is
00
fu(u) = f fx(v)fy(u-v)dv
Proof. We consider another random variable V to obtain the new
continuous bivatiate (U, V) where U = X + Y and V = X .
Then u = x + y, v = x (1)
. theJacobian
..
a(u.v)
8(x,y)
=:av :av =1:~I=-l
8x By
8(u, v) .
:. -- has same sign for all (x, y)
8(x,y)
fu.v(u, v) = fx.Y(X'Y)I~~::~~1
00
= J f x (v) fy (u - v) dv using ( 1)
(1)
(v-mx)2 + (u-v-my)2
Now,
()"2 ()"2
x y
= (v-mx)2 +{u-(mx+lny)-(v-mx)}2
()"2 ()"2
x y
J
BIVARIATE DISTRIBUTIONS (CONTINUOUS VARIATE) 273
_ ax2 2 axay
[Putting f1.=mx+-2 (u-m) and A=--]
a a
(u-mf , (v-JJ)2
1 --,-
= __ l_e- 20-2 xl -:f(v)=--e 20-- IS tspdf
&0' A&
(u-m)2
I - 20-2
=--e
&0'
U is a normal variate with parameter (m,a)
Theorem 3. If X and Yare independent Gamma variates with
parameters (/, 1) and (m,l) respectively then
(i) (X + Y) is a Gamma variate with parameter (l + m,l)
X J-l
(ii) The pdf of - is fv(v) = I ,v> 0 where v=X
Y (l+v) +mB(l,m) Y
EM-2A-IS
ENGINEERiNG MATIlEMATlCS-lIA
274
X
Proof. Put U =X + Y and V = -Y
. x uv u
i.e. U = x +y and v = -. From these we get x
Y
= --and
l+v
y = --.
l+v
C(U,V)jW al
ll
1 1 x 1 x+y
Now, C(x,y) : ~ = =- I ~y=-7
ax: 0'
which is always negative since x,y > 0 for Gamma dist.
= fx (x)fy (y) --
l
x+y
=e
-x I-I
x .e y
-y m-I
l ": X, Yare Gamma variate
r(1) rem) x+y
e-(x+Y)xHym-1 .L
r(l)r(m) x +Y
«: (_uv Y- I
(_u \,"-1 2
1+ v) u 1 + v)
=--~--~~~~-- --~---~
f(l)r(m) (1 + (~+V)2
l -r v
»:\
1+v)
e -u "ul-l+m-I+2 • vi-I 1+ v
f(l)r(m)(1 + v)H+m-I+2 U (1 + v)
-u I+m-I /-1
e u v
= f(l)r(m)(1 + vi+m ; otherwise fu,v (u, v) =0
The pdf of U is
00
o eo -u I+m-I I-I
e u v
=
--«>
J Odv+ J r(l)f'(m)(1 + v)'+m dv
0
-u I+m-I f'(l)f'( )
=e u for u z-t) .:B(/,m)= m
r(/+m) r(/+m)
which is the pdf of Gamma variate with parameter (l + m, 1) .
•
:. X + Y is a Gamma variate with parameter (l + m, 1) .
The pdf of V is
co
o eo -u I+m-I I-I
e u v
=
--«>
f Odu+
0
J f'(/)f'(m)(1 + v)'+m
du
V
I-I
f
<Xl
-u I+m-Id
= f'(/)f'(m)(1 + v)l+m 0 e u u
. vH
= -Tt] + m)
f'(l)f'(m)(1 + v)l+m
lI
V- ru + m)
(1 + v)l+m r(l)r(m)
Vl-I 1
(1+v)'+m B(l,m)
=0 otherwise
ENGINEERING MATHEMATICS - II
276
Now, J J f(x,y)dxdy =I
--«J--«J
2 2-x
or, J J C(x + y)dydx =1
x=Oy=O
12 I
? 1 R y=;
or, C·- J(4- x-)dx =1
20 2
(detail calculation is not shown)
or, C =_.3
8
Now, P(X <l,Y >~)= p(o<X < l,~< Y < 2-X)
12-x 31 2-x
=J JC(x+y)dydx=-J Jex+y)dydx
o 1 80 1
"2 y="2
35
= 64 (detail calculation is not shown)
:.fx(x)=e-x, x;:::O
o 0 0 R
J
= e -x e -xdx =
o 0 --:-¥-------x
4
= _![e-2XJoo = J.. ',: lim e-2x =0
2 0 2 x--+oo
:.Jx(x)=2(l-x),O<x<l
y=X
,.q-q
Y
and Jy(Y) = J 2dx=2y, O<y<1 0
x=o
(ii) Conditional density of X given Y = y, 0<y <1
JYIx (y / x) = J(x,y)
Jx(x)
= 2
2(l-X)
= -,
I-x
1
x <y <1
~
= ffx/y(X/'!:)dX=l/f~dx=~
1/2 13 1/2 1_ _ 4
3
Example 4. Determine the value of k which makes
or, J
x=o y=o
1 k x y dx dy=Y or, k J X[y2]X
x=o 2 0
dx=l
k 1 3 k 1
or, - f
x dx = 1 or _. - = 1 :. k = 8
, 2 4
2.<=0
The marginal pdf of X,
00
fx(x) = J f(x,y)dy
-00
x
= f kxydy, Gc xc l
y=o
X2
=8x·-=4x3
2
:. fx (x) = 4x 3
, 0<x <1
=0 ,elsewhere.
00
"* j(x,y)
:. X and Yare not independent.
Example 5. The joint pdf of (X, Y) is given by
1
j(x,y)=-(6-x-y), O<x<2, 2<y<4
8
= 0
, elsewhere
Find P(X < I,Y < 3), P(X + Y < 3), the marginal distribution,
=-
1
J (6---
3 1
y)dy=- J --
1 (11 3 )
Y dy=-
3
8 y=2 2 8 y=2 2 8
BIVARIATE DISTRIBUTIONS (CONTINUOUS VARIATE) 281
I 3-x 1 I I [ 2 ]3-X
= I I g(6-x- y)dxdy =- I 6y-xy_L dx
x=Oy=2 ' 8 x=O 2 2
1
=- II[{ 6(3-x)-x(3-x)--2- (3-X)2} - { 12-2x-2 }] 1 5 =-5
dx =-x-
8 x=o 8 3 24
00
:·fx(x)=
4iI g(6-x-y)dy=g l[ 6Y-XY-T
i]4
y=2 - 2
1 1
=-(6-2x)=-(3-x),O<x<2
8 4
00
21 1 1
.',fy(Y) = I-(6-x- y)dx =-(10-2y) =-(5 - y)
08 8 4
The conditional density of X,
O<x<2
Now, P(X<1/Y=3)=P(O<X<1/Y=3)
I 16-x-3 II
= ffx/y(xj3)dx = f dx=- f(3-x)dx
o 0 2(5 - 3) 40
P(O<X<I,2<Y<3)
= by definition of conditional probability
P(2 < Y <3)
3 I
3 I 1
f f f(x,y)dy dx f f -(6-x- y)dxdy
y=2x=O y=2x=O 8
= "-----3 ----
3 1
f/y(y)dy f-(5 - y)
2 24
Example 6. If f(x,y) = 3x2 - 8xy + 6y2, 0 < x < 1,0 < y < 1 is the joint
pdf of the bivariate (X, Y) then find the conditional densities fy (xl y ) and
fy (y Ix ). Hence find whether X and Yare independent.
co
Solution. Now the marginal pdj, fx (x) = f f(x,y)dy
I
f (x/y)_f(x,y) O<x<1
x - Jy(y) ,
3x2 -8xy+6y2
or, fx(X/Y)~ 2 ,O<x<l,givenanyy,O<y<l
1-4y+6y
,
BIVARIATE DISTRIBUTIONS (CONTINUOUS VARIATE) 283
p (~ < X <
4 4
l/r = ~)
2'
Y
fx(x) = J f(x, y) dy o x
x
= J 2 dy for 0 < x < 1and for an x, 0 < y < x
y=O
=2x
:. fx(x) = 2x, O<x<1
00 I
Now,fy(y)= J f(x,y)dx= J2dx for O<y<1andforany,y<x<1
--co y
= 2(1- y)
:·fy(y)=2(I-y),O<y<1
The conditional density of X given Y = Yo' 0 < Yo < 1,
_f(x,yo) 2 1
f xtr (Ix Yo ) - + ( ) = = -- in R for Yo, Yo < x < 1
JY Yo 2(l-yo) I-yo
3/ y=- 1J = J fx/y(x/~)dx
(1 3/4
P -<X<-
4 4 2 1/4
3/4 1 1
= J Idx using (1) where Yo ="2
1/21- -
2
=2(~-~J=~
Example 8. Abivariate distribution for X and Yare given by the pdf
f(x,y) = e",
0 < x < y < 00
=0 , otherwise.
Find the conditional density of X given Y =Y.
Solution. the region Y
R:{O<X<Y<oo} y
R .
x
is open. It is shown
in the adjacent figure.
In this region, 0 < Y < 00
and for any y, 0 <x <y
__--~~-----------x
The marginal density,
ao Y
fy(y) = J f(x,y)dx = J e-Y dx = e-Y [x1~
--ao x=o
=ye-Y, O<y<oo
Required conditional density, for 0 < x < y,
l.e., fx/y(x/y)=!.., uc x c v .
y
Example 9. The pdf of the random variable X is fx (x) =11, 0 S x S 1.
1
Solution. fy/x (y/x) = -, 0 <y <x <1
x
=0 , elsewhere
r (/x) = fx,y(x,y)
We know, )Y/X y fx(x)
1
or, fx,y(x,y) = fy/x (y/x)fx(x) =~'1, 0 <Y <x <1
:. From (2),
fu(u) == f fu.v(u, v) dv ,
v=o .~.
when 0 < u < 1
/I
Now, x=v, v
y=-.
u
j
BIVARIATE DISTRIBUTIONS (CONTINUOUS VARIATE) 287
Obviously as 0 < x < 00, 0 < y < 00, 0 < v < 00, 0 < u < 00
ox ox
o(x,y) - 0 1 v
Now, o(u,v) = ou 8v = v 1 = u2
8y oy
u2 u
au ov
Obviously, o(x,y) is always positive.
o(u,v)
By theorem, fv v(u, v)
.
= fx .
y(X,Y)\O(X,Y)\
o(u, v)
-a(v+~) V
= a2e U._
u2'
0 < u < 00
,
0 < v < 00
:. the pdfof U =X is
Y
fv(u) = 1
-00
f(u,v)dv= la2e-a(v+~J .-;-forO<u<oo
·~o u
a2 -{I-J)V
Ie
00
=2 ".vdv
u 0
a-7
= -;;
{
(0 - 0) +
1.
1
00
Ie" (I)}
-a 1+- v
dv .,' lim~=O
. a(l+~)O X~OO eX
288 ENGINEERING MATHEMATICS - lIA
-Jl+~)V
""
a2 U e ul /I
~7Ta(u+1) [ -a([+~) 0
J
( + ~1)
= a (0-1) .: lim ~=O
x-+oo eX
-u(u + l)a I .
1
=---
(u+ Ii
:. the density of U =~ is fuCu) = (u ~ 1)2' 0 < u < 00.
:. 8Cu, v)
8Cx,y)
=1' 11=-1
1 0
which is always negative
=_1 .1=_1
x2 v2
. 1
Thus the joint pdf, fu,v(u, v) = -? ' v z I, 0 < u - v < 1
v-
eo
1
or, fu (u) = 1- -, 1~ u ~ 2
u
1 1
=----, 2~u<oo
u -1 u
Exercise 8
1. The joint distribution of X and Y is given by the pdf
1
f(x,y) = -, (x,y) E R
2
=0 , elsewhere
where R is the interior of the triangle having vertices (0, 0), (2, 0), (1, 2).
Find P(X s 1, Y s 1).
I 2.< 1 I I 1
[Hint: Required probability = f f - dy dx + f f - dy dx ]
Oy=O 2 ~y=o 2
2
2. The joint distribution of (X, Y) is given by
f(x,y) = k, (x,y) E R
= 0 , elsewhere
=0 , elsewhere
find (i) k (ii) marginal density of X and Y (iii) P(X + Y ~ 3) (iv) conditional
density of X given Y = y. (v) Whether X and Yare jndependent,
I 3-.<1
[Hint. (iii) P(X + Y s 3) = f f -(2x + 5y)dydx]
oy=2108
EM-2A-19
E GlNEERING MATHEMATlCS-IlA
290
\ \
1) \ x-2 2\ 1
[Hint. (iv) p IX - yl s - ~ I I (x + y)dydx + II (x + y)dydx
~ 2 ~ 0 O x+I
5. f(X,y)=sinXSiny,O<x<~, 2 O<y<1r 2
=0 ' elsewhere
is a joint pdf of (X, Y). Find the marginal density of X and Y. Prove that
f(x,y)=2 , x+y<l,x>O,y>O
= 0 , otherwise
Find conditional density function of Y given X =x .
7. The joint pdf of (X, Y) is given by
f(x,y)=kx, O<y<x<l
= 0 , elsewhere.
Find k, Find the marginal density function of X and Y. Are X, Y
independent ?
BIVARlATE DISTRIBUTIONS (CONTINUOUS VARIATE) 291
f(x,y)=6(l-x-y), x>O,y>O, x+ y c I
=0 , elsewhere.
Find the marginal distribution of X and Y. Are they independent?
11. The joint pdf of (X, Y) is
f(x,y) = 3(X2y + y2X), G c x c l, O<y<1
=0 , otherwise.
(I) Find the marginal density function (ii) Find p(.!.3 < Y s ~/.!.
3 2
< X ~~)
4
12. Raindrops fall at random on a square R with vertices (1, 0), (0,1),
(-1, 0), (0, -1). If (x, y) is the point struck by a random and X, Yassume
the value x and y respectively having joint distribution given by the pdf
f(x,y) =.!. if (x,y) E R
2
=0 , otherwise
fmd the marginal pdf of X and Y. he X and Y independent?
I-x 1 x+1 1
[Hint.fx(x)= J -dy,
22-
O<x<land fx(x) = J -dy, -1<x~O]
y=x-I y=-x-I .
of X + y.
15. The pdf of two independent random variables are respectively
fx(x) =~, - 2 < x < 2 and fy(Y) =~, - 2 < Y < 2 respectively. Find the,
distribution of X + Y . [Hint. Put V = X + Y, V = X ]
16. The pdf of two independent random variables are respectively
17. The pdf of two independent random variables X and Yare given
by fx(x) =1, O~x~l andfy(y)=l, O~y~l
...) 37/
(III (iv) not independent
/648
4. (i) 1
(Hi) X + ~ , 0 < y < 1, 0 < X < 1 and X + ~ , 0 < x < 1, for 0 < y < 1
x+i Y+i
3
(iv) "4 (v) dependent
5. fx(x)=sinx,o<x<~, fy(y)=siny,O<y< ~
1
6. --, O<y<l-x
I-x
1 13
7. 28' 168; not independent
3((1- 2)
8. 3,3x2 (0 < X < 1), y, 0 < y < 1, not independent
2
13. (i) 3
3 3
(ii) fx =i x2 + x, 0:S; x:s; 1, fy =i / + y, O:S; y:S; 1
(iii) 311/3456
. _ 6y(x+ y) _ 6x(x+ y) < <
(IV) fx/y - ,0 :S; y :S; 1 fx/y - ,0 - x_I
2 +3x ' 3y+ 2
294 ENGINEERING MATHEMATICS-IIA
~(66+76)=71.
x = _1 (2 x 3 + 4 x 2 + 1 x 4 + 3 x 1)
10
_ 1 I
or , X = -( 6 + 8 + 4 + 3) = -
x 21 = 2.1
10 10
Theorem 1. If the two variables X and yare related by the equation
x-c. _ x-c .
Y =d tnen Y = d 'wherec and d are any number,
Proof We consider the frequency distribution of X :
X: Total
y: YI ... Yn Total
... In N
300 ENGINEERING MATHEMATICS-IIA
X· -c
where Yj = _,_ .
d
1 1 x. -c
Now the A.M ofy, Y = N L/;
II
1=1
Yj = N
II
L/;
1=1
"s:
=-I(/;Xj
1 II - /;c) = - 1 ("I/;Xj - I/;c
")
Nd j=1 Nd j=1 j=1
1 n C 1 1 11 C 1 c x-c
=-L/;Xj --=--I/;Xi -- =-x--=--.
Nd j=1 d d N j=1 d d d d
Note. The above theorem is very much helpful to determine the A.M of
the variable assuming large data. This is shown in the following Illustration.
Illustration. We are given the following Grouped frequency distribution:
Weight in gmsfx]: 110-119 120-129 130-139 140-149150-159
Frequency: 5 7 12 20 16
Weight in gms(x): 160-169 170-179 180-189
Frequency : 10 7 3
To fmd the A.M we construct the following table:
Calculation of Mean
Class- interval Midpoint Frequency x, -154.5
Xi-1545 Lv.
10
(Xj) (Ji) =»
110-119 114.5 5 -40 -4 -20
120-129 124.5 7 -30 -3 -21
130-139 134.5 12 -20 -2 -24
140-149 144.5 20 -10 -1 -20
150-159 154.5 16 0 0 0
160-169 164.5 10 10 1 10
170-179 174.5 7 20 2 14
180-189 184.5 3 30 3 9
Total - 80 - - -52
BASIC STATISTICS 301
8
Here N = 80 and 'LhYi = -52 .So the
i=1
. Xi -1545 _ x -1545
Smce Yi = 10 ' so Y = 10
x-1545 -
or, -0.65 = or, X = 148.
10
Thus the required A.M, x = 148.
Theorem. 2. (On Mean of Composite Group)
Let XJ,X2,.· .... x,. be the A.M. of r groups containig nl ,n2,.····· n,
observations respectively. Then their combined mean or composite mean
is given by
Proof Omitted.
..; Mode = 1m + fi - 10 X i
2/1 - 10 - 12
where 1m = lower class boundary of the modal class
II = frequency of the modal class
10 = frequency of the class preceeding the modal class
12 = frequency of the class succeeding the modal class
i = width of each class (note that this is same for each class).
EM-2A-20
ENGINEERING MATHEMATICS -llA
306
Illustration.
Consider the grouped frequency distribution.
Marks No. of Candidates
____ --I
0-9 4
10-19 9
20-29 12
30-39 18
40-49 20
50-59 12
60-69 10
70-79 9
80-89 4
90-99 2
Here we see every class has same width which is 10 and only class
40-49 has heighest frequency. So the modal class is 40-49. As we know
11-/0
Mode = I + //I xi
2/1 - 10 - 12
where 11/1 = lower class boundary of the class 40-49 = 39 ·5
11= frequency of the class "40-49"=20.
10 = frequency of the class preceeding the class 40-49 = 18
12 = frequency of the class succeeding 40-49 = 12
i = width of each class = 10 .
:. Mode =39.5+ 20-18 xl0=41·5.
40-18 -12
Relation among Mean, Median and Mode: For a distribution having
single Mode the relation is Mean - Mode == 3(Mean - Median)
Note. For a symetrical distribution mean, median and mode coincide.
Measure of Dispersion
9.6. Variance and Standard Deviation.
As we have stated in the previous article A.M represents the entire set
of datas.But the degree to which the datas tend to spread about the A.M
is to be measured.It is usually measured by variance or standard deviation
which are discussed below:
J
BASIC STATISTICS 307
Variance.
The mean of the squares of the differences of the observations (or
datas) assumed by a variable from their arithmetic mean (A.M) is called
variance of the variable.
Standard Deviation.
The positive square root of variance is called Standard Deviation (s.d)
Thus (i) if XI> X2' .... x; be the datas then their variance,
. . crx
deviation, = + - 1 ~~ J1"(i Xi - X_)2 .
N ;=1
If the datas have grouped frequency distribution then Xi will be the
mid -value of each class-interval and Ji would be the corresponding
frequency.
Illustration: Let X be a variable which assumes the datas:
20,85,120,60,40.Then x=.!.(20+85+120+60+40)=65.To find the
5
variance and standard deviation we go through the following table:
Xi x;-65
20 -45 2025
85 20 400
120 55 3025
60 -5 25
40 -25 625
Total 6100
308 ENGINEERING MATHEMATICS-IIA
5 2
Here :L(Xj - 65) = 6100
j;)
1 5 2 1
Then Var(x) = - :L(Xj - 65) = - x 6100 = 1220.
5 j;) 5
and the S.d, c = .J1220 = 34.93.
also.There
I"
Var(x)=-IJ;x?- where N=j.,+j2+···+j,,·
(In)2
-IJ;Xj
Nj;) Nj;)
Theorem 3. If the two variables x and yare related by the equation
7 7
Here, N = 10, IhYi = -2 and Ihyl = 114
i=1 i=1
1
Now, Var(y) = N IhY/ - ( N1 IhYi )2 = 751 xl27- (27)2
75 =1·56.
1
310 ENGINEERING MATHEMATICS -IIA
ay 2 1
y=-=-xl00=-xlOO=2- which is lesser than that of A. We may
y 80 2
conclude batsman B is more consistant i.e, reliable than A.
9.8. Moments. The r-th moment of the values (datas) xl'x2,' ... ·x"
about a number A is defined as
.!.{(XI - At + (x2 - AY + +(xn - AY} =.!. f,(x; - At
n n~
If the datas have the frequency shown in the following table
datas (x) XI x2 x3 ,xn
frequency (J;): 1; 12 J; In
then their r-th moment about A is
n
where ]V = 1; + 12 + ... + In = 'LJ;
;=1
BASIC STATISTICS 311
= 17168 ·17
x 2 4 3 Total
3 2 4 10
=0·912
The third raw moment of x
Theorems
(1) 1st central moment =0
1" _ 1
1st central moment =- 'L(x; -x)/; =N 'Lx;.!; -x'LJ;
(n _n)
N ;=1 1=1 1=1
1 n { - _}
= Nf;:J; (x, _A)2 -2(x-A)(x; -A)+(x-A)2
,In '2 - 1 1 - 2 n
=-'L/;(x;-A) -2(x-A)-'LJ;(x;-A) +-(x-A) 'L/;
N ;=1 N N ;=1
, - , - 2
= m2 - 2(x- A)ml +(X - A) (1) ':'LJ;=N
- 1 n 1 11
I" In In ,
=-'L/;x;--'LJ;A =-'LJ;(x;-A)=ml
N ;=1 N;=I N ;=1
x-c
Theorem. If two variables x and yare related as y = -- where c,
d
d are constants then m (y) mr(x) =
dr r
mr(x) and mr(y) are central moments ofx andy respectively
- x-c
.
Proof. We know, y =--
d
x-c x-c x-x
:.y-y=-----=--
d d d
- x·-x
:.y. -y=_'- d
I
for i=12' , 3, .
1 n 1 n {x.-c
~_c}r
Now, m,.(Y)=NLJ;(y;-yY
1=1
_
=N~J;
1-1
=r=r
=~ I,J; (x; -rxY
N ;=1 d
1 1 n -
= -dr' N ~J; (x; - x)" ': d' is independent of i
1=1
1 mr(x)
or, mr(y) = dr mr(x)=-----;r-
9.11. Skewness and Kurtosis
The skewness and kurtosis of a frequency distribution are
y\ (skewness)
m
=--+(J
where m3, m4 are 3rd and 4th central moments and (J is the standard
deviation of the distribution.
Note: Since skewness and kurtosis are ratio of two quantities having
same unit so they have not unit, they are pure number.
Example. Following is the frequency distribution of a variable x :
x : 112·45117·45 122·45127·45 l32·45 l37·45 142·45 Total
f : 5 15 20 35 10 10 5 100
Find its skewness and kurtosis
314 ENGINEERING MATHEMATICS-IIA
x.-127·45
x; J; s, -127·45 ' 5 =y; J;y; J;y/ J;y/ J;y/
;-~-3. ~l-(-~)+2(-~J
1 33 2 201
= -- + - - - =- = 0·804
2 25 125 250
,
-I 315
BASIC STATISTICS
I m4(Y) = m4' (Y),- 4m3' (y)ml' (y) + 6m2' (y){ml' (y)f - 3{ml' (y)f
I
1 1
= 6 2-4{ -~}( -l)+6. 5 {-lJ -3(-lJ
5
= 62 _~+ 66 __ 3_= 782"'-=12.5232
5 10 125 625 625
. ~-127·45
Smce y- =
5'I
. m4(x) 7827
and KurtOSISof X'Y2(X) =---34 =-,--3 = -0·316
ux 54-
9.12. Significance of Skewness
Skewness shows the extent of symmetry of the frequency diagram
of a variable. Below we draw the frequency diagram of the variables
x,y,z and w respectively.
Frequency diagram of x
J;
__ ~4-I'
XI
__ --~-+--------------------+~
x2 X3 x . Xs x6 x7 x9
4 Xg
Note that this frequency diagram is symmetric about the frequency
Is . This type of frequency distribution will have skewness O. The curve
1is fitted to show the symmetry.
3]6 ENGINEERING MATHEMATICS-lIA
Frequency diagram of y
----~~~~~~~~~~~====~Yi
y, Y2 YJ Y4 Y7 Y9
Note that this diagram is not symmetric. It has positive value of
skewness as this has asymmetry towards +ve direction of X axis. The
curve f is fitted to show this asymmetry.
Frequency diagram of z
Ii
Note this diagram also has asymmetry towards right side. Its
asymmetry is more than that of y. Its measure of skewness will be greater
than that of y. It is also positively skewed.
Frequency diagram of w
BASIC STATISTICS 317
Note that this frequency diagram has asymmetry towards left side. Its
skewness will be negative.
9.13. Other Formula For finding Skewness
Skewness can also be measured with the help of following theorems
whose proofs are kept beyond the scope of the book.
Mean -Mode
Theorem 1. Skewness = -------
Standard deviation
3 (Mean - Median)
Th eorem. 2 Sk ewness = --'-------'-
Standard deviation
9.14. Significance of Kurtosis
Kurtosis shows the peakedness of the frequency diagram of a variable.
Below we draw the frequency diagram of the variables x, y, z and w
respectively
Frequency diagram of x
•I
318 ENGINEERING MATHEMATICS-IIA
x. -37
/
class mid point frequency y; = 15 I.».
interval (x;) /;
25-29 27 16 -2 -32
30-34 32 28 -1 -28
35-39 37 14 0 0
40-44 42 12 1 12
Total 70 -48
I __
But y=x--
37
5
1 i.e., x=37+5y=37- 24 x5 =33.57.
35
So the total wage of 70 workers = Rs. 33·57 x 70 = Rs. 2349·9.
Thus the wage of total 80 workers is
Rs (2349·9 + 600) = Rs 2949·9.
r 2949·9
., Mean wage is Rs 80 = Rs 36·87.
,I Example. 2. TheA.M calculatedfrom the following frequency distribution
I is known to be 72.5. Find the value of x :
Xi -74·5
Classes Midpt Frequency yJi
Yi= 10
(Xi) (/;)
f 3~39 34.5 2 -4 -8
40-49 44.5 3 -3 -9
5~59 54.5 11 -2 -22
6~9 64.5 20 -1 -20
7~79 74.5 x 0 0
8~89 84.5 25 1 25
9~99 94.5 7 2 14
Total 67+x -20
- LYi/; -20
.. Y = L/; = 67 + X
.. X = 74.5+10Y= 74.5-~
67+x·
320 ENGINEERING MATHEMATICS-IIA
1
BASIC STATISTICS 321
or, X 144k = 48 .
=
3k
Example. 8. Find the median of the following frequency distribution:
x 5 10 15 20 25 30 35 40
f 7 10 15 18 23 21 17 8
The cumulative frequency distribution table is given below:
x f c.]'
5 7 7
10 10 17
15 15 32
20 18, 50
25 23 73 ~
30 21 94
35 17 111
40 8 119
EM-2A-21
322 ENGINEERlNG :\1ATIIEMATICS-lIA
Here N = 119
:.
N + 1 = 60
2 N+1
So the cumulative frequency just greater than -2- is 73 and the
value of x corresponding to c.f 1JTs-z5. Rence tlie median is 25.
Example. 9. Calculate mean, median and hence find the approximate
value of the mode from the following frequency distributions:
'Height (inches) : 60-63 64-67 68-71 72-75 7fr-79 80-83
No of students 8 3 18 6 16 8
To find mean we first construct the following table :
x·-73·5
classes Midpt frequency cf Yi = 4
I yJi
(Xi) (f;)
60-63 61.5 8 8 -3 -24
64-67 65.5 3 11 -2 -6
Xi -73.5
Now, Yi = 4 or, Xi = 73.5 + 4Yi
N_F
:. Median = 1m + 2 xi = 715 + 29.5 - 29 x 4 = 71.83 .
J; 6
Using the relation Mean-Mode = 3 (Mean-Median) we have
:. Mode = 70.6596
Example. 10. Calculate the mode of the following data:
1, 12, 5, 8, 12, 13, 8, 1, 4, 8, 7, 8, 5.
Let us arrange the given variaties with corresponding frequencies as
given below
x f
1 2
4 1
5 2
7 1
8 4
12 2
l3 1
As the variate 8 ocurs 4 times which is maximum, so,
mode = 8.
Example. 11. Calculate the mode from the following distribution:
class : 10-15 15-20 20-25 25-30 30-35
frequency 6 9 11 7 7
Here the greatest frequency 11 lies in the class 20-25. Hence modal
class IS 20-25.
.. 1m = lower class boundary of the modal class
=20·
fl = frequency of the class = 11
fo = frequency of the proceeding the modal class
=9·
f2 = frequency of the class succeding the modal class = 7 .
i = width of each class =6.
324 ENGINEERING MATHEMATlCS-IIA
- I +
~-J;o. X I = 20 +
11-9
x 6
.. Mo d e - m
J
2f.. - fo - f2 22 _ 9 -7 = 22·
Ex.12. The median and mode of the fo 110wing frequency distribution are
known to be 27 and 26 respectively. Find the values of u and /3 :
class interval 0-10 10-20 20-30 30-40 40-50
frequency 3 u 20 12
Since mode = 26, so it lies within the class 20-:-30.
Mode = I + fi - fo x i = 20 + 20 - u x 10
In 2fJ - fo - f2 40 - u -12
= 20+ 20-u x 10
28-u
10(20 - u) _10--,-(
2_0_-_u-,-)
=6
20+ = 26 or,
28-u 28-u
u=8.
N_F
Now median = I + 2 xi
In fm
43 + /3 -11
21 + /3
.. 27 = 20 + 2 x 10 or --
'4 = 7 or 21
,t-'
+ (.l. = 28 /3 = 7.
20
BASIC STATISTICS 325
x, -53
class Midpt frequency r, = 15 /;Yi /;Y/
interval (Xi) h
36-40 38 . 14 -3 -42 126
41-45 43 26 -2 -52 104
46-50 48 40 -1 -40 40
51-55 53 33 0 0 0
56-60 58 50 1 50 50
61-65 63 37 2 74 148
66-70 68 25 3 75 225
Total 225 65 693
.. Var(y) = N
1 "LhYi 2 - (1N "LhYi )2
1
= 225 x 693 -
( 65
225
)2 = 2.996 .
2 2
Then the correct LX/ = 297642·8 - 85 + 58 = 293781· 8 .
2293781·8 2
Hence the correct (s.d) = 70 -(64.61) =22·43·
This shows that the daily wage of the workers of factory C is most
consistant.
Variability is highest for factory A.
Example 16. Calculate the first four central moments, the skewness,
kurtosis for the following distribution of 1083 cases of entric fever.
BASIC STATISTICS 327
1
328 ENGINEERING MATHEMATICS - I1A
-3(-2·721)4
= 55·408
S·mce y. = xi-32·5
I 5
:. the central moments of x are given by m, (x) =0
m2(y)=m~~x) :.~(x)=25x3.781=94.525
m 4 (y) __m4(x)
54 or, m4 ()x = 62 5 x 55· 408 = 346 30· 00
m3 (X) 829·875
Skewness of x, Yl(x) = --3- = 3 = 0·903
(Yx (9·722)
. m4(x) 34630
Kurtosis of x r2(X)=--4--3= 4 -3=0·876
, (Y x (9· 722) .
I Example 17. The first four moments of a distribution about 5 are 2, 10,
40 and 218. Find the first four central moments and moments about O.
I Solution. Let m, and m; are central moments and moments about 5
respectively.
\
Using the relation between Inr and we have m;
I m2 = m2 "2- m1 = 10 - 22 = 6
I ,
m3 = m3 - 3m2 m1
"
+ 2ml '3
= 40 - 3 x lOx 2 + 2 x 2 =-4
3
m4 = m4 , - 4m3 "6"2
Inl + m2 m1 -
3Inl '4
I = 218 - 4 x 40 x 2 + 6 x lOx 22 - 3 X 24 = 90
I which are first four central moments.
1/1 I" - - In - 1 -
:.- L(x; _a)2 =- L(x; _X)2 +2(x-a)·- L(x; -x)+-n(x-ai
n ;=1 n ;=1 n ;=1 n
332 ENGINEERING MATHEMATICS -IIA
In - - 1 1 - -
=- Lex; _X)2 + 2(x-a){- LX; --. nx} + (x _a)2
n ;=1 n n
In -2 - - - - 2
=- L(x; -X) +2(x-a)(x-x)+(x-a)
n ;=1
or, 2nd order moment about a
27. Find the s.d and variance of the variable x having following
frequency distribution :
x 94.5 84.5 74.5 64.5 54.5 44.5 34.5
f· 2 12 22 20 14 4 1
28. Find the mean and s.d of the following distribution:
Class-interval : 4-6 6-8 8-10 10-l2 12-14 14-16
Frequency 13 111 182 105 19 7
29. Find the standard deviation of the set of numbers in the arithmetic
progresion 4, 10, 16, 22, 28, ... 154.
30. Find the standard deviation for the frequency distribution of life
time (in hours) of 80 light bulbs:
Life-hours:500-600 600-700700-800 800-900 900-1000 1000-1100
Frequency: 6 12 25 18 14 5
31. The mean and standard deviation of marks of 70 students were
found to be 65 and 5.2 respectively. Later it was detected that the value
85 was recorded wrongly and therefore it was removed from the data
set. Then find the mean and s.d for the remaining 69 students.
32. Following are the maximum daily temperature (0 c) recorded in
a week in Kolkata :
38,40,36,35,30,32,34.
Using transformation property determine the s.d of maximum daily
temperaturae in Farenheit scale.
41. The mean and s.d of the income of teachers in two schools are
given below:
Mean income s.d
School A 8000 1200
School B 9000 1350
Compare the variability of the incomes of the teachers in two schools.
42. Following are the scores of two batsmen, A and B in ten innings:
A : 84 73 115 36 19 7 12 6 119 29
B : 13 42 12 48 51 4 47 76 37 0
Who is the more consistent player.
43. For a class of students the height has a distribution with mean
162 em, s.d 10 em and weight has mean 57 kg, s.d 8 kg . Compare' the
variability aspect of the distribution of height and weight.
44. Thew scores of two batsman A and B in 10 innings are
19 31 48 53 67 90 10 62 40 80
A
32 28 47 63 71 39 10 60 96 14
B
Find which batsman is more consistent in scoring.
Type I 310 9
Type II 260 14
compare the relative variability of life of the type of bulbs.
46. You are given the distribution of wages in two factories X and Y.
Wages (Rs): 50-100 100-150 150-200 200-250 250-300 300-350
No. of X 6 11 18 32 27 11
workers Y 2 9 29 54 11 5
State in which factory the wages are more variable.
BASIC STATISTICS 339
47. Find the first three central moments of the following frequency
distribution:
Yearly income 3-6 6-9 9-12 12-15 15-18 18-21 21-24 Total
in lakh
Nu of'families 28 292 389 212 59 18 2 1000
48. ·Find the first four central moments and the skewness, kurtosis
for the variable x having the following frequency distribution:
x 21-24 25-28 29-32 33-36 37-40 41-44
f 40 90 190 110 50 20
49. Find mean, mode and standard deviation for the body weight of
the children having following frequency distribution. Hence find the; .
measure of skewness :
Body weight : 14·5 15·516·5 17·5 18·519·5 20·5 21·5
No. of children: 35 40 48 100 125 87 43 22
50. Find the skewness of the following distribution
Wages (x) 55-58 58-61 61-64 64-157 67-70
No. of workers: 12 17 23 18 11
51. Find the skewness and kurtosis for the following distribution:
x 4.5 14.5 24.5 34.5 44.5 54.5 64.5 74.5 84.5 94.5
f 1 5 12 22 17 9 4 3 1 1
52. From the following distribution, find the skewness and explain its
significance.
Weekly: 70-80 80·90 90-100 100-110 110-120 120-130 l30-140 140-150
wages (in Rs)
No. of: 12 18 35 42 50 45 20 8
employees
53. (a) Find skewnes based on mean-median from the following
frequency distribution.
Marks No. of Students 'Marks No. of Students
Above 0 100 Above 50 50
" 10 98 " 60 35
" 20 95 " 70 23
" 30 90 " 80 13
" 40 80 " 90 5
340 ENGINEERING MATHEMATICS- IlA
x-x
z = -- a , where a is S.D of x, then show a
r
= mr jar .
59. The first three moments of a distribution about 2 are 1, 22 and
10. Find its mean, s.d and the skewness.
60. The first three moments of a distribution about 2 are respectively
1, 16 and -40. Prove that the mean is 3, the variance is 15 and the third
central moment is -86. also find the first three moments about O.
61. The variance of a symmetrical distribution is 25. What must be
the value of the fourth central moments so that the distribution is (i) lepto
kurtic (ii) mesokurtic (iii) platykurtic.
62. Calculate the first four central moments for the following data and
examine for the nature of the distribution:
x 1 2 3 4 5 6 7 8 9
f 1 6 13 25 30 22 9 5 2
63. Find the second, third and fourth moments about mean of the
following frequency distribution. Hence find the skewness and kurtosis
of the distribution.
BASIC STATISTICS 341
1
/
71. The mean, median and the coefficient of variation of the weekly
wages of a group of workers are respectively Rs 45, Rs 42 and 40. Find
the (i) mode "(ii) variance (iii) skewness, for this distribution.
72. The distribution of the Wages of the Govt. employee in a country
is such that its first two moments about 4 are respectively 1.5 and 2.7.
The median of this distribution is 2.1. What will be the shape of the
frequency diagram of the distribution.
73. For a group of 10 items, LX; = 452, LX; = 24,270 and the Mode
= 43·7. Find the skewness.
74. For a moderately skewed distribution, mean is 160, mode is 157
and standard deviation is 50.
" Find (i) coefficient of variation
(ii) skewness
(iii) median.
75. In a distribution, Mean = 65 , Median = 70 and skewness = -0·6 .
Find Mode and the coefficient of variation.
76. Given mean = 50, C.v = 40010, skewness = -0.4. Find s.d, Mode
and Median.
77. Consider the following distribution :
Distribution of x Distribution of y
Mean 100 90
Median 90 80
s.d 10 10
Find whether (i) distribution of x has same degree of the variation as
distribution of y. (ii) both the distribution have the same degree of
skewness.
'. 78. The mean of a certain distribution is 50, its s.d is 15 and skewness
is -1. Find the median.
79. The skewness of a distribution is 0.32. Its standard deviation and
mean are respectively 6·5 and 29. 6. Find the mode and median.
80. In a frequency distribution the c.v = 5, s.d =2 and skewness
= 0 .5 . Find the mean and mode of the distribution.
BASIC STATISTICS 343
Answers
l. 7·6 2. 5·7 3.80 4. 3 ·98 5.501 6.715 7. 150
8.62·16 9.79 10. 1: 2;50,100 11. 2·1 12. 76, 38 13. 30
2n +1 n +1 4n2 -1
14. (i) -3- (ii) -2- (ill)-3- (iv) n + 1(v) 2n(n + 1)2
. 3n(n+l)
(VI) 2(2n+l)
15.127,6; 130 16.60 17. Both are 6·5 18. (i) 2280
19. 10 20. 250, 150 22. 65 23. 1 24. (i) 1982·40
(ii) mean = 1765, median = 1533·33, mode = 1070
L
345
BASIC STATISTICS
Xi
~1 0 3 4
/; 3 2 1 4
1" 1" 2
(c) - ~Xi -- ~Xi (d) None.
n i n i
9. The standard derivationof the followingobservation is 5, 7, 1,2,6,3.
(a) 4.66 (b) 2.16
(c) 1.47 (d) none.
10. The mode of the following data is : 2, 1,3,2,1,5,2,2,1,6,4,21,3
(a) 5 (b) 2
(c) 3 (d) 1.
11. The median of the following distribution is: 7,9,5,3,10, 15,21,
19, 17
(a) 15 (b) 9
(c) 10 (d) 17.
12. The median of the following distribution is: 10, 13,9, 7, 37, 16,27,32
(a) 16 (b) 14.5
(c) 13 (d) 15.5.
346 ENGINEERING MATHEMATICS -IIA
21. Two set of datas {Xi} and {Yi} are related by Yi ==12xi - 3115. If
the median of the first set is 130 then the median of the second set is
(a) 124.5 (b) 130.5
(c) 140.5 (d) none of these.
22. The mode of the frequency distribution
X 0 1 2 3
f 8 24 36 10
is
(a) 0 (b) 1
(c) 2 (d) 3.
23. The mode of the frequency distribution
n 0 1 234
f 23 24 21 24 20
are
(a) 0 (b) 1
(c) 2 (d) 3.
24. The AM of 100 observation is 2.5. So the AM of 50 of these
observation is more than 2.5
(a) True (b) False.
25. The median of 100 observation is 2.5. So 50 of these observation is
2.5 or more
(a) True (b) False.
26. If you remove the largest data from a group of different datas, the
AM of the remaining datas always changes
(a) True (b) False.
27. If the relation between the two group of observations {Xi} and
28. If the relation between two set of observations {Xi} and {Yi} is
Xi =2 Yi + 5 and median of X is 25 then the median of y is
(a) 20 (b) 10
(c) 12.5 (d) none of these.
29. If the relations between two set of observations {Xi} and {Yi} is
2Yi - 6xi = 6 and mode of the 1st set is 21 then the mode of the second set
IS
(a) 13 (b) 29
(c) 55 (d) none of these.
30. The number of observations of two groups of datas are in the ratio
2: 1 and their A.M are 8 and 128 respectively then the A.M of the combined
group is
1
(a) 88 (b) 45-
3
(c) 48 (d) none of these.
31. The mean of the observations 1,2,3,. =.n with frequencies
12,22,32,. ··,n2 respectively is
3n(n + 1) n(n + 1)
(a) 2(2n + 1) (b) 2(2n+l)
3n(n + 1)
(c) 2n + 1 (d) none of these
34. The datas of the two groups {x j} and {Yj} are related by
,
Xj -800 . .
Yj = 50 and If s.d of the second group IS 2.6257 then the s.d of the
first group is
(a) 131.29 (b) 135.16
(c) 134 (d) none of these.
35. The s.d of maximum daily temperatures in centrigrade scale is .
3.16. Then the s.d of those of in Farenheit scale is
(a) 5 (b) 7.1 (c) 5.69 (d) 6.69
C F-32
[Hints: 5= 9 is the relation between Centrigrade and Farenheit
scale).
36. 2xj + Yj = 3 is the relation between two sets of datas {xj} and
{Yj} . If c x = 3 then o y =
(a) 3 (b) 4 (c) 6 (d) none of these.
37. The standard deviation of the datas -5, -10, -12, -19, -20 is a
positive number
(a) True (b) False.
38. The s.d of n number of observations xpx2,. ",xn be s then the s.d
of -xt>-x2, •• ,-xn is -s
(a) True (b) False.
39. The variance of first n natural numbers is
n2 -1
(a) n2 -1 (b)--
10
n2 n2 -1
(c) - (d) --
12 12
40. The variance of the frequency distribution
x 1 2 3 n
I, 1 2 3 n is
(n + 2)( n - 1) (n + 2)( n + 1)
(a) 18 (b) 18
Answers
1.b 2.d 3. a 4. b 5.c 6. a 7. c
10.1 Introduction:
In an earlier chapter we introduced Bivariate data and its distribution.
Consequently the correlation between two variable was discussed. In basic
statistics we are concerned with only the bivariate data and the correlation
- regression of these. In order to go through an independent study of this
chapter we again give the idea of bivariate data with an example. It is
very obvious to mention that there may exist two or more random variables
on an event space. Two random variables show two characters of the
outcomes of the experiment. In this chapter we are concerned of two
random variables defined on an event space.
10.2 Bivariate Data
Let X Y be two random variables defined on an event space S of an
experiment. Then the pair (X Y) is called a Bivariate on the same event
space. Such a bi-variate (X y) assumes a pair of rea Is (x, y) corresponding
to an outcome in the event space where x and yare assumed by the random
variables X and Y respectively against the same outcome. These pair of
values (x, y) are called Bi-variate Data.
Illustration: Let us consider the experiment of drawing a student from
the colleges of West Bengal. Then the event space S = Set of all college
students ofW. B. Let X and Ybe two random variables defined on S such
that X ( a student) = his / her height and Y (a student) = his / her weight.
Then the pair (X Y) is a bi-variate defmed on S ; a pair of values say
(5.6, 63.4) means X assumes the value 5.6 and Yassumes 63.4 against a
particular student, i.e. the height and the weight of a particular college
student is 5.6 and 63.4 respectively. (5.6, 63.4) is a particular bi-variate
data ; only 5.6 or only 63.4 is an uni-variate data. So, if
x Xtl
y Yl YII
then (x, y): (Xl' YJ), (X2' Y2), ... "', (XII' Ytl) are the bi-variate data.
CORRELATION, REGRESSION RANK CORRELATION 353
Note: For a bi-variate (X; Y) assuming the datas (x, y) there mayor
~ may not have any relation or association between x and y. For example if
x = height and Y = weight of a student then there may exists a relation
or association between x and y. But if x = height and Y = I.Q of a student
then so far as we know there is no relation or association between x and
y. In the subsequent sections of this chapter we will be mainly concerned
with the character' association'.
10.3. Scatter Diagram
Thedigramrnatic representation of bivariate data is scatter diagram.
Let (XIoYl), (X2,Y2), ... "', (x",Yn) be nnumber of bivariate datas.
For each (Xi' Yi) we get a point P; on the X- Y plane whose co-ordinate
is (Xi' ») as shown in the figure. .
y"
•
Y2 •
Yl
:X
0 Xl x2 X"
....
•
.. . . •
--+-------------------~x
o Fig!
If the scatter diagram is like Fig 2 then we say the association between
the two variables x and y is negative, i.e. if x increases y has tendency to
decrease.
y
o 0
o
o
·0 0
o
o • 0
--I-------------------+x
o Fig 2
If the scatter diagram is like Fig 3 we say the association between the
two vairables x and y is 0, i.e. there is no association between the two
variables x andy.
y
. ... .. .
o •
o •
• 0 0
o •
~~-----------------+x
o Fig 3
If the scatter diagram is like Fig 4 then we say the two variables x
and yare perfectly positively associated, i.e. if x increases y also increases
- never decreases at any stage. [WB.V.T. 2005
CORRELATION, REGRESSION RANK CORRELATION 355
-o~--------------~x
Fig 4
If the scatter diagram is like Fig 5 then we say the two variables x
and yare perfectly negatively associated, i.e. if x increases y decreases -
- never increases at any stage. [W.E. U.Tech 2005 ]
Y
•
o x
Fig 5
Illustration.
Suppose the bivariate datas assumed by the bivariate (X; Y) are given
below:
x : 1 1.8 2 2.5 3 3.5 3.8 4.4 5 5.5
y :.9 1.6 2.1 2.1 2.4 2 3 3.1 3 3.5 3.8
We plot the points (1, .9), (1.8,1.6), (1, 2.1) . and so on. Thus the
scatter diagram of these data is shown below :
y
6
5
4
...
3
•
2
1 •
2 3 4 5 6
356 ENGINEERING MATHEMATICS - IIA
From the scatter diagram we see x and yare positively associated but
not perfectly. As x increases y has a tendency to increase; at some stage
it decreases.
10.4. Correlation and its Different Types.
Let us have a bivariate data (x, y). The degree of association or the
strength of relationship between the two variables x and y is called the
correlation between the two random variables X and Yor the correlation
between their assumed datas x and y.
If Y has a tendency to increase as x increases we say x and yare
positively correlated. That is if the scatter diagram is like Fig 1 of the
previous Article then we say x and yare positively correlated.
If y has a tendency to decrease as x increases we say x and yare
negatively correlated. If the scatter diagram is like Fig 2 of previous Art
then x and yare negatively correlated.
If the value of yare not affected by the changes in the values of x
then we say x and yare uncorrelated or zero correlated.
Illustrations.
Let (X; Y) and (Z, W) be two bi-variates where the datas assumed by
them are
x 1 2 3 4 5 6
y 2.5 1.5 3.5 2.5 3.5 1.5
and
z 1 2 3 4 5 6
w 2.5 3 4.5 4 5.6 6
From the two sets of datas we see the degree of association between x
and y is lesser than that between z and w. This would be more clear if we
draw the scatter diagram of them. In this case we say 'correlation of x and
y' is lesser than the 'correlation of z and w '.
Y : YI Y2 Y3 Yn
Then the covariance of X and Y is
n ;:1
i
Cov(x, y) = J.. (x; - x)(y; - y) where x and yare the Arithmetic
Cov(x,y) =-
1 n
LX;Y; - -Lx;
(1 n )( 1 n
- LY;
)
n ;:1 n ;:1 n ;:1
Proof: By definition
= -1 L..Jx;y;
~( _ _
- x;y - xY; + x Y
__ )
n ;:1
= -1"
L..Jx;Y;-
_1" _1" __ 1"1
Y- L..Jx;- x- L..JY;+ x Y- L..J
n n n n
1" 1
=- L..Jx;Y;- Y x -x Y +x y-·n
n n
5
we have Lx;=-2-1+0+1+2=0
;=1
5
LY; =4+1+0+1+4=10
;=1
5
LX;Y; =-2x4+(-1)xl+OxO+lxl+2x4
;=1
= -8 - 1+ 1+ 8 = 0 .
Therefore their covariance,
=~XO-(~XO)(~XIO)=O.
Correlation Coefficient.
Let (x, y) be a bivariate data assumed by a bi-variate (X; Y). Then the
correlation coefficient of x and Y is denoted by
(To deal with big values we need the help of the subsequent theorems)
CORRELATION, REGRESSION RANK CORRELATION 359
x y xy X2 y2
-6 -4 24 36 16
-4 -3 12 16 9
-3 -1 3 9 1
-1 - 1 1 1 1
1 0 0 1 0
2 2 4 4 4
4 3 12 16 9
7 4 28 49 16
0 0 84 l32 56
cri=~LX?-(~LX;r =~XI32-(~XOr 3;
crxcry
(2) Since Cov( x, y) and c x' cry have same unit so rxy is a pure number.
rt \
360 ENGINEERING MATHEMATICS -IIA
(3) Below we give three scatter diagram of the bivariate data (x, y) against
which different sign of rxy is shown :
y
y
..
..
X X
o Here rxy > 0 0 Here rxy < 0
y
..... ......
...
.
. . ..
--+-----------------~~
o Here r 0 =
xy
In 2 In , .
=- :La2(X; - X) = a2 - :L(x; - xt = a2 Var(x)
n ;=1 n ;=1
CORRELATION, REGRESSION RANK CORRELATION 361
Similarly o , = I c 10" y
1 II
=!fac(x; - x)(y; - y)
n j;\
1 n
= ac- 'L(Xj - x)(y; - y) = ac Cov(x, y)
n j;\
Cov(u, v) ac Cov(x,y)
Therefore, ruv
O"uO"v lalO"x\cIO"y
ac Cov(x, y) ac
or, ruv = lal \cl 0" xO"y = lall~ rxy .
Theorem 3. Let (x, y) and (u, v) represent two sets of bivariates datas
such that u = ax + b and v = cy + d . If a, c have same sign then ruv = rxy .
Proof: This is deduced directly from the previous Theorem 2.
There we get .
ac
ruv = lal \cl rxy .
Theorem 4. Let (x, y) and (u, v) represent two sets of bivariate datas
such that u = ax + b and v = cy + d . If a, e have opposite sign then
ae
Proof: From Theorem 2 we deduce ruv = lallel 1"..y.
If a is positive and c negative then lal = a, Icl = -c. So from above
ac
we have ruu = -( -) rxy = -r xy .
a -e
ac
Similarly if a < 0 and e> 0, ruv = -(-)- rxy = -r xy .
-a c
Theorem 5. If (x, y) represents bi-variate data then x + y and x - yare
two variates. Then
Var(x + y) = cri + cr~ + 2cr xcr yrxy
Var(x - y) = C"i + cr~ - 2crxcr yrxy
where o stands for s.d and rxy stands for correlation coefficient.
Proof: Var(x + y) = -
1
LII {
(Xi + ») - X + Y
-----}2
n i=l
1 ~{
= - L... Xi
_
+ Yi - X - Y
_}2 [ ..
n i=l
1 II 2
=- L {(Xi - x) + (Yi - y)}
n i=l
1 " 1 " 2 II
Corollary: (l) If the two variates x and yare uncorrelated then rxy = 0
and hence
Var(x ai + a~,= Var(x - y)
+ y) =
(2) Var(x + y) = ai + a~ + 2Cov(x,y)
Var(x- y) = ai + a~ -2Cov(x,y)
Theorem 6. Let x and y be two variables whose means are X, y; s.d
U=--
x-x v=--
y-y
ax' ay·
Then we have, from previous theorem,
Now, !tul=!t(x;_X)2
n ;=\ n ;=\ ax
1 1 " 2 1
=-- 2 "(x.
~
-x) I
= -·a
2
2
x
=1 (2)
....a x n ;=\ ax
Similarly, !t v? = 1 (3)
n ;=\
(u; + v;Y~ ~ 0
or, UT +2u;v; +VT ~O
n II n
or, Lul + L2u;v; +L vl ~ 0
;=\ ;=\ ;=\
or -
,
lL21L
U +- I 2u·v·
I I +-1L2 v· >0, -
n n n
ENGINEERING MATHEMATICS -lIA
364
2
or, I+-Lu;v;+I~O from (3)
n '
1 n
... (4)
or, - LU;v; ~-I
n ;=1
_ x-x - y- Y
Now, U =--=0 and v =--=0
crx cry
1
we get - L U; v; ~ 1 and consequently rxy ~ 1.
n
Thus -1 s rxy s 1.
Theorem 8. If rxy = i1 (the max and min value of- correlation
coeffiecient) then y is a linear function of x and vice-versa.
x-x - y-y - [W.B. U. Tech. 2008
Proof: Let u=--, v=--.
o, cry
First part: Let rxy = il
=2i 2 Cov( U, v)
If rxy = 1 then take '-' sign and if rxy = 1 then take + sign in (1).
'J
CORRELATION, REGRESSION RANK CORRELATION 365
l'a a
Then rxy = Illla(xx =~ rxx (1)
x =!( -2 -1 + 0 + 1+ 2) = 0
5
Y =!( 4 + 1+ 0 + 1+ 4) = 2
+-----~~~----~x
o
5
. Cov(x,y) 0
•• rxy = =--=0
<Jx<J y <Jx<J y
Illustrative Examples
Example 1. Find the correlation coefficient between the two variables x
and y where the bivariate datas are given by :
x 65 66 67 67 68 69 70 72
y 67 68 65 68 72 72 69 71
Since the datas are so large we subtract 67 from x and 68 from y to
make the datas small.
Calculation of Correlation Coefficient.
Xi v. Xi -67=Ui Yi -68=Vi 2
u, Vi
2
u.v,
65 67 -2 -1 4 1 2
66 68 -1 0 1 0 0
67 65 0 -3 0 9 0
67 68 0 0 0 0 0
68 72 1 4 1 16 4
69 72 2 4 4 16 8
70 69 3 1 9 1 3
72 71 5 3 25 9 15
Total - 8 8 44 52 32
= i (i )(i
x 32 - x 8 x 8) = 3
av
2 = -;I LVi 2 - (I-; L )2 = "81 x 52 - ( "81 x 8)2 = 55
Vi
Example 2. How would the scatter diagram be look like if rxy = -I.
In a previous theorem we have seen if rxy = -I the relation between x
.!l+ ~T)')ci cx
or, k =- cr xcr J:: rxy) = cry .
= 104 -5x4=i
5 5
c x2 I
=-"X· 2 (I
n L. - -"X·
n L.
I
)2
I
I
=-x650-
25
(-x125
1
25
)2 =1
436 324 36
---=-
25 25 25
EM-2A-24
370
10.6. Regression
From the bivariate datas given by (x, y) the estimation or prediction of
the average value of a variable, say y , corresponding to a specified value
of the other variable x is called regression. The average value of x may
also be predicted corresponding to a specified value of y.
Illustration.
Below we are given pairs of marks in Mathematics and Mechanical
Science obtained by students in an Engineering College.
x (Marks in Math) : 65 66 67 67 68 69 70 72
y (Marks in Mech Sc.) : 67 68 65 68 72 72 69 71
Depending on these data let we are to predict the marks in Mechanical
Science obtained by a student who scored 73 in Mathematics.
There may exist many students scoring 73 in Mathematics. Every suet
student may not score same marks in Mechanical Science. In the proces:
of regression the average of these marks is predicted.
If this average is 75 we say the predicted value ofy is 75 correspondiru
to x = 73.
CORRELATION, REGRESSION RANK CORRELATION 371
---o~----------~--------------------~x
3.6
______________ ~ jL ~_
372
ENGINEERING MATHEMATICS - IU
Let (XI>Yl), (X2, Y2), (X3, Y3)··· ... (xn, Yn) be n pairs of observation
given by the bivariate data (x, y). Let i, y be the means; CJ.o CJ
y be th
standard deviations of the values of x and Y respectively; rxy is correlatio
coefficient.
Then the equation of the (i) Regression line of Y on x is
Y- Y--b
- yx h
(x-x -) were byx =r - CJy
xv
. CJ
x
(ii) Regression line of x on Y is x - i = bxy (y - y) , where bxy = rxy . CJ x .
CJ I'
Proof: (i) Let Y = a + bx (1)
II n
n n n
or, LXiYi = a LXi + b LX?
i=l i=l i=l
n n n
or, LXiYi =(y -bi)Lxi +bLxl, putting the value of a from (3)
i=l i=l i=l
CORRELATIO ,REGRESSION RANK CORRELATION 373
1" __
- ~XiYi -x Y Cov(x,y)
or, b = -'-'n'--- _
1" 2 -2 cri
-~Xi -x
n
y = y - bx + bx or, y - y = b(x - x)
c
or, Y - Y = rxy· cry (x - r) which is the required 'regression line ofy on x'.
x
Regression Coefficients
. cry
Th e quantity bl'x = rxy - appeared in the 'regression equation of y
. crx
on x' is called regression coefficient of yon x.
Illustration.
Let us consider the bivariate datas given by (x, y) as
x - 6 - 4 -3 -1 1 2 4 7
Y - 4 -3 -1 -1 ° 2 3 4
1 11 1
Here we see X=-LXi =-(-6-4-3-1+1+2+4+7)=0
n i=1 8
21
2
The correlation coefficient,
av J7 fl4
byx = rxy .-- = 0.97 x r;:;;:; = 0.97 x r;:;;:; = 0.632·
ax v33 v33
.fi
Therefore in these observations the equation of the regression line of
yon x is
y - 0 = 0.632(x -0) or, Y = 0.632 x.
Now if we are to predict the value of y when x = 2.3 then it is
y (predicted) = 0.632 x 2.3 = 1.454 .
. If there exist many values of y corresponding to x = 2.3 then
approximately 1.454 is average of all of them.
10.8 Properties of Regression Line and Coefficients.
The properties of regression line and the regression coefficients are
presented as the following theorems.
Theorem 1. For a bivariate data given by (x, y) the correlation coefficient
and the regression coefficients 'have same sign.
ay
s;
£ . Pr Q .. • s::e the regression coefficients are given by
b.; -
same sign as
.
1xv -
- ay
rry.
and since P x' a yare
= r.; cr.,
always positive byx, bxy are of
CORRELATION, REGRESSION RANK CORRELATION 375
Proof: 'Since byx = rxy cry and bxy = rxy ~ therefore byx x bxy = (rXy)2 .
crx cry
Theorem 3. The two regression lines intersect at the point (x, y).
Proof: The equations of the two regression lines are y - y = byx (x - x)
and x - x = bxy(y - y) .
We see both the equations are satisfied by x = x, y = y . Hence proved.
Theorem 4. The two regression coefficients have same sign.
Proof: Since byx x bxy = (rxy)2 = positive ;.the result is proved.
'Cov(x, y) . Cov(x, y)
byx = 2 and bxy = 2
crx cry
91 71 1 1 1 1
97 75 7 5 49 25
108 69 18 -1 324 1
121 97 31 27 961 729
67 70 -23 o 529 o
124 91 34 21 1156 441
51 39 -39 -31 1521 961
73 61 -17 -9 289 81
111 80 21 10 441 100
57 47 -33 -23 1089 529
10
- -LUi
10
2
( ) =-x6360-(0)
10
2
=636
I
CORRELATION, REGRESSION RANK CORRELATION 377
2
aV=-L>i
1 2- (1-LVi )2 1
=-x2868-0
2 =286.8
10 10 10
1x l
and ruv = NNrxy = rxy :. rxy = 0.913 .
a
So the regression coefficient of y on x, byx = -..l'...rxy.
ax
.J286i
= ../636 (Note that if rxy was not required we could find byx by direct
applying Theorem 6)
and the regression coefficient of x on y,
ax ../636 .
bxy = -rxy = ~ x 0.913 = 1.36
ay v286.8
:. the regression line of y on x is
y-y=by,r{x-x) or, y-70=0.61(x-90)
'---
or, y = 0.61x + 15.1 (1)
and the regression line of x on y is x - x = bxy (y - y)
or, x-90=1.36(y-70) or, x=1.36y-5.2 (2)
When x = 100, Yesttmated = 0.61 x 100+ 15.1= 76.1, unit [from (1) ]
We got rxy
2
= -9 .', rxy = ±_3 .
16 4
Since we know the regression coefficient byx and r
xy
have same sign
so r-,:y is negative as byx is negative.
To estimate the value of x we use the regression line 'of x on y '. So,
from (2) we get for y=1.5, 4x+9xI5+5=O.
185
or, 4x=-185 :. xesl =--=-4.625
4
Example 3. If x = 4y + 5 and y = kx + 4 be two regression equations
of 'x on y' and of 'yon x' respectively then find the interval in which
k lies. [ W.B. U. Tech 2004 ]
Since x=4y+5 is 'ofx ony' and since its gradient is 4"I :. I =4"1
bxy
or, bxy=4.
CORRELATION, REGRESSION RANK CORRELATION 379
1 7 1 5 - 1- 7 - 1- 5
U=-X+- v=-y-- .. U=-X+- V=-Y--
2 4' 3 2 2 4' 3 2
2 cry 2 2
Now, =--rxy =-byx =-x 3= 2 .Ex.
3 crx 3 3
Example 5. The relationship between travel expenses (y) and the duration
of travel (x) is found to be linear. A summary of datas for 102 pairs is
given below:
(i) Find the two regression coefficients (ii) Find the two regression
equations
(iii) A given trip has to take seven days. How much money should a ,
salesman be allowed so that he will not run short of money ?
_ 1
- I" 1
x=-L..x=-x51 0=5 y=-x 7140= 70
n 102 ' 102
Cov(x, y) = 188.24 = 12
byx = 2 86 (by Th. 6.)
c, 15.6
(ii) The regression line of y on x is y - y = byx (x - r)
or, y-70= 12(x-5)
or, y = 12x+ 10
and x - x = ro x (y - y) .
Oy
cry _ rcry
tan8=+ rcrx crx
1+ cry .rcry
rc , o,
. 2 2
= ± _1_-_r_. cry . _-::-cr--,x,--~
2 2
r o, o, + cry
1- r2 cxcry
=+-------"--
- r cr/+cr/'
. As r2 ~ 1 and o x .o yare both positive, so the positive sign gives the
acute angle between the lines.
l-r2 crxcry
Hence tan8 = 2 2
r crx +cry
7t
When r = 8,8 ="2' So in this case the two required line are
r
xy
2 =(-~)(-~)
4 3 = ~12 < 1 .
Thus our hypothesis is correct.
Hence equation (1) is the regression equation of y on x and equation
(2) is the regression equation of x on y.
Let x,y be the mean of x and y, respectively
.. From (1) and (2), we have.
3x + 12y = 19
3Y+9x=46.
Solving there equation we get
- 5 -
x= ,y=3".1
Again we have
2 1
r xy = 12
1
r
xy
=±--
2..fj .
But we know that the regression coefficient
byx and y have same sign, so r xy must be nagative as byx is negative.
So, the required correlation coeffcient is
1
r
xy
=---
2..fj
Also we have
CORRELATION, REGRESSION RANK CORRELATION 383
1
= - 2./3
1
4
crx:cry=2:./3.
-=)=----
n -n
where d is the difference of corresponding rank in x and y and n is the
number of individuals.
Proof. Beyond the scope of the book.
Example. Calculate rank correlation between the performance of 8
students in Physics and Chemistry.
384
ENGINEERING MATHEMATICS -IIA
4 6 4
-2
1 7 -6 36
6 8 4
-2
Total - - 110= Ld2
o!
6Ld2
:. rank correlation =1- 3 =1- 6xll0 =1- 660 =-0.3095
n -n 83 -8 504
10.11. Rank Correlation when there is Tie Rank in any series
\I
If there exist more than one individual with same rank in either or both
n
the series we say there exists Tie Rank.
If k number of individuals get Tie Rank 'r' then the Average Rank of
After assigning this average rank to each of the k individuals, the next
rank assigned will be r + k and so on. ar
C(
E
CORRELATION, REGRESSION RANK CORRELATION 385
Example:
Individual 1 2 3 4 5 6 7 8 9
Rank x 5 3 5 1 2 5 4 6 2
Ranky 3 1 6 4 3 2 5 7 8
In series x, we see the individuals 5 and 9 obtain tie rank 2. So each of
them will be assigned the average rank 2 + 3 = 2·5 . Next the individual 2
2
will be ranked 4, each of the individuall, 3, 6 will be assigned average rank
Rank Correlation = 1-
6{Ld2+L~}
3 12
n -n
where d is the difference of the corresponding rank and t denotes the
number of individuals having the tie- rank in the first or second series.
Example. We consider the previous example :
Individual 1 2 3 4 5 6 7 8 9
Rank x 5 3 5 1 2 5 4 6 2
Ranky 3 1 6 4 3 2 5 7 8
Here some individuals get tie-rank. We assign the average rank to them
and the revised ranking is shown in the following table. Hence the Rank
correlation is determined.
EM-2A-2S
386 ENGINEERING MATHEMATICS -IIA
(R,) (~)
1 5 3 7 3·5 3·5 12·25
2 3 1 4 1 3 9
3 5 6 7 7 0 0
4 1 4 1 5 4 16
5 2 3 2·5 3·5 1 1
6 5 2 7 2 5 25
7 4 5 5 6 1 1
8 6 7 9 8 1 1
9 2 8 2·5 9 6·5 42·25
1
CORRELATION, REGRESSION RANK CORRELATION 387
Solution. Here n = 12
Calculation of Rank Correlation
36 51 7 8 1 1
98 91 1 1 0 0
25 60 9 6 3 9
75 68 4 4 0 0
82 62 3 5 2 4
92 86 2 2 0 0
62 58 6 7 1 1
65 35 5 10 5 25
35 49 8 9 1 1
n -n
or, O·5 = 1-
6x Ld 2
or, Ld 2
= 82 . 5
103 -10
:. before correction Ld' = 82 . 5
After corrcetion, Ld 2
=82.5-32 +72 =122·5
/
CORRELATION, REGRESSION RANK CORRELATIO 389
Solution. Let R), R2' R3 be the ranks given by the three judges respectively.
Calculation of Rank Correlation
=1- 6Id;
3
=1- 6x214 =1_1284 =-0.2969
n - n 103 -10 990
We see the Rank correlation between judgements of judges I and ill is
maximum. So judges I and ill have nearest approach to common teste of
beauty.
t 3}- 1
= 1-
6
{Ld + L-U
2
= 1-
6{4l+3}
= 0 . 733
3
n -n 103 -10
Exercise 10
[I] Short Answers Questions
1. If the relation between variables X and y, U and V are
2X + 3Y = 4,3U + 4V = 5 and the regression coefficient of X on U is 4,
fmd the regression coefficient of Y on V .
2. For a set of bivariate data x and y, the lines of regression are
4x + 3y+ 7 = 0 and 2x + 5y= 4. Identify the lines and hence find the
correlation coefficient between x and y.
3. The bivariate data (x, y) results the following:
LXiYi = 414, LXi = 120, LYi = 90, LXT = 600, LYT = 300.
Calculate the correlation coefficient between x and Y if the number of
datas is 30. [ W.B. U. Tech 2005 ]
4. The results of a bivariate analysis are given in the following form :
16.Given that x = 45, o, = 2.5, Y = 60, cry = 3.2 and the correlation
coefficient 0.75. Find
LXiYi = 414; LXi = 120, LYi = 90, LX? = 600, LY? = 300, n = 30
If z = 5 + 3Y obtain the estimate of z when x = 15 .
20.First is of x on y; 2nd is of y on x.
15. The following results were obtained for 25 pair of bivariate datas :
x==5, y=4, crx=2, cr~==2.25, rxy==0.6
It was later detected at the time of checking that one pair of values
x = 5, Y = 4
was copied wrongly in computing. the above results. Find the
exact correlation coefficient between x and y after correction.
16. The following results relate to bivariate data on (x, u):
2:>jUj = 414 ; 2:>j = 120 ; 2:>j ==90, Lxr ==600 Lur ==300 n = 30
30
y : 12 14 18 25 31
Predict y when x = 15.
311 •
23. The following are the scores of 10 students studied for a
two mathematics test and their scores on the test :
the Hour of study (x): 4 9 10 14 4 7 12 22 1 17
was Test score (y) : 31 58 65 73 37 44 60 91 21 84
'ere Find the equation of least square line apprroximate of the test score
tion on the numbers of hour studied. Also predict the average test score of a
student who studied 14 hours for the test.
[ Hint: 'leas square line' means regression line. Find both.]
398 ENGINEERING MATHEMATICS -UA
28.Iftwo regression lines are 3x+9y =46and 3y+ 12x = 19 find which
one is the regression line of x on y. If the variance of x is 4, fmd the means
of x and y, the variance of y, and the correlation coefficient between x and
y.
29.If the equations of the two regression lines are 3x + 12y = 19 and
3y + 9x = 46, determine which one of these is the regression equation of
y on x and which one is that of x on y. Give reasons for your answer.
Find also the arithmetic mean, correlation coefficient and ratio of variances
ofx andy.
30.If y = -12x and x = -0.6y are two regression lines, compute rxyand
cr2 x
-2·
cry
x+ y : 24 28 30 33 35.
32. Find the rank correlation coefficient for the following data :
(i) x 80 91 99 71 61 81 70 59
y 123 l35 154 110 105 l34 121 106
(ii) x 78 89 97 69 59 79 68 57
y 125 137 156 112 107 136 123 108
[Hint. first assign Rank to each of the serie x and y]
(iii) x 75 88 95 70 60 80 81 50
y 120 l34 150 115 110 140 142 100
33. The marks obtained by the students in Physics and Chemistry are
as follow:
Marks in Physics 30 33 45 23 8 49 12 4 31
Marks in Chemistry : 35 23 47 17 10 43 9 6 28
Compute their ranks in the two subject and find the rank correlation
of the two series of marks. Give your interpretation of the result.
Candidates :A B C D E F G H
Judge I :5 2 8 1 4 6 3 7
Judge IT :4 5 7 3 2 8 1 6
37.The rank often students in Statistics and Mathematics are as follow
Statistics 6 4 9 8 1 2 3 10 5 7
Math. 3 5 8 4 7
1 6 10
9 2
Find the rank correlation coefficient. Give your interpretation.
Rice 75 88 95 70 60 80 81 50
Wheat 120 134 150 115 110 140 142 100
Compute the rank correlation of production of Rice and Wheat in
different countries.
39. Two series of values are given as
X 130 132 128 130 127 125
y 115 134 120 130 124 128
Find the Rank correlation coefficient between the two series of values.
40. Ten competitors in a beauty contest are ranked by three judges in
the following order :
Judge I 1 5 4 8 9 6 10 7 3 2
Judge II 4 8 7 6 5 9 10 3 2 1
Judge nr 6 7 8 1 5 10 9 2 3 4
Use rank correlation to measure which pair of judges have the nearest
approach to common taste in beauty.
41. Following are the I.Q and Scores of ten students in a class:
Student 1 2 3 4 5 6 7 8 9 10
I.Q 100 100 110 140 150 130 100 120 140 110
Score 35 40 25 55 85 90 65 55 45 50
Find the rank correlation coefficient between I.Q and Scores of
students.
CORRELATION, REGRESSION RANK CORRELATION 401
42. Obtain the Rank correlation coefficient from the following two series
of observation :
X : 62 58 68 45 81 60 68 48 50 70
Y : 68 64 75 50 64 80 75 40 55 64
43. The following table gives the monthly family income in thousand
rupees of 11 students and their scores in English. Find the rank correlation
coefficient between monthly family income and scores in English:
Income 40 46 54 60 70 80 82 85 85 90 95
Scores 45 45 50 43 40 75 55 72 65 42 70
44. Compute the correlation coefficient of the following ranks of a
group of students in two examination. What conclusion do you draw from
the result?
Roll Nos. I 2 3 4 5 6 7 8 9 10
Rank inH.S
Exam. 5 8 6 7 4 2 3 9 10
RankinJEE
Exam 2 5 7 6 3 4 8 10 9
45. Ten competitors in 'a table tenis tournament are. ranked by three
judges X, Y, Z in the following order:
Rank by X 1 6 5 10 3 2 4 9 7 8
RankbyY 3 5 8 4 7 10 2 1 6 9
Rank by Z 6 4 . 9 8 1 2 3 10 5 7
Discus which two of these three judges has the nearest approach to
common assessment.
Answers
2. Positively correlated 3. 0.977 4. 0.5533 5. 0.8961 6.0.7027
7. rxy = 0 ; there is relation like y = x2. 8. 0.0096, almost no association.
EM-2A-26
402 ENGINEERING MATHEMATICS - IIA
26. The second equation is 'of y and x' ; the first is 'of x on y
'x-=l,y-=-l. ./3
,rxy =--.
2
2
27.
3
28. 3y+12x=19 .
IS 'f 0 xony ' ; x=-,
- 1 y=5,Var
- ()y =-,rxy=-
16 1r;;
3 3 2v3
29. 1st is ofy on x, 2nd is ofx ony.
- - 1 1 2 2
x=5, y="3' r=- 2./3 crx: cry=4: 3.
30. - 0.85, 0.5 31. 0.98 32. (i) 0·952 (ii) 0·952 (iii) 0·929
33. 0·9; If Rank in Physics is higher then there is a tendency that the
rank in Chemistry would be higher also
2
34. -0·357 35·0·84 36."3 37. -0·3 38.0·939.0·33
(d) none.
CORRELATION, REGRESSION RANK CORRELATION 403
cov(x,y)
(c) a -a (d) O. [W.E.U.Tech 2007]
x y
(d) r«.
8. The regression coefficient of x on y is given by
cov(x,y) cov(x,y)
(b) 2 (c) 2 (d) none.
ax ay
9. byx x bxy; where bxy ,byx and r are regression and correlation
coefficient, is
(a) r (b) r2
1 1
(c) - (d)2' [WE.U.Tech,2006]
r r
ENGINEERING MATHEMATICS -IIA
404
1
13. If u == _ x + 2, v == 5x + 1 and r llv
1
== -, then rxy == [c
1
3 5
1 (d) 3.
1 (c) -
(b) - 5
(a) 5 3 then
and x,y are uncorrelated
14. If var(x) == 25, var(y) == 9
var(x + y) == (d) none.
(b) 34 (c) 16
(a) 8
x-X y-y -
15. If u==_,v ==-, then rxy-
o, cry
(d) none.
(b) cov(x,y) (c) cov(u,v)
. 4 1
16. For given datas, we have var(x) == 1,var(y) =="9 and rxy =="5.
Then var(x - y) ==
77 (d) none.
13 53
(c) 45
(a) - (b) 45
9
CORRELATION, REGRESSION RANK CORRELATION 405
(c) a parabola
1 1 1
(a) 18 (b) --
18 (c) 20 (d) none of these.
•
• •
then rry is
(a) 0 (b) -1
(c) negative (d) positive.
28. If the two bivariates (x,y) and (u, v) are such that x = -2u + 4
and y = 3u - 6 then
(d) b = d = O·
30. If (x,y) is a bivariate where var(x)=0.2,var(y)=0.1 and
correlation coefficient of (x,y) is 0.01 then variance of x + y is
approximately
(a) 2.300 (b) 0.3 (c) 0.3028 (d) 0.5069.
31. For the bivariate (x,y) s.d of x + y is same as the s.d of x - y if
(a) x andy are negatively correlated
(b) x and yare positively correlated
(c) x and yare uncorrelated
(d) x and yare perfectly correlated.
32. If var(x) = 0.2, var(y) = 0.09 and covariance of x,y is 0.1 then
the s.d of x+ y is
(a) 0.3 (b) 0.7 (c) 0.11 (d) 0.1.
408 ENGINEERING MATHEMATICS-IIA
33. Two variables x andy are such that CJx = I,CJy = 2,CJx_y = J2 then
cov(x,y) =
1 2 3
(a) - (b)- (c) 1 (d) 2.
2 5
. 34. Two variables x and yare related as 3x + 4y =5 then correlation
coeffieient of x and y is
1
(a) -1 (b) 1 (c) 2 (d) O.
35. If the two variable x andy are related as 2y2 +5x =9 then rxy =1
(a) True (b) False.
36. Correlation coefficient of a variable with itself is 1
(a) True (b) False.
37. If two variables are independent then they are uncorrelated
(a) 1 (b) 2
(c) -1 (d) none of these.
53. If the mean of the variate x and yare respectively 5 and 70 and the
reqression coefficient of x on y is 0.08 then the regression line of x on y is
(a) y=12x+l0 (b) x=8y-0.6
1t
(a) - (b) tan-I.!..
4 7
(c) tan " 7 (d) none of these.
55. The two regression lines become identical. The correlation coefficient is
(a) ±1 (b) 0 (c) only 1 (d) none of these.
11.1 Introduction:
Consider the bivariate (X, Y) assuming the data (xPYj). We consic
a scatter diagram of (X, Y) as follow :
.. . i
o 10 11 12
In the scatter diagram we see there are many dots having co-ordin
(XI' Yj) correspending to X = XI . The ordinates of these dots are YJ
corresponding to x;::: XI . Let m y (x I) be the mean of these Yj. 1
bold dot on the line of X = XI represent this my (XI) i.e. the ordinate
this bold dot is m/x l). Similarly the bold dots on the line
X = x2,x = x3' X = x4 ..... are found which represent the means 'of
corresponding to X = x2,x = x3'x = x4 .
In the figure the smooth well known curve Y = ¢(x) is drawn in su
way that though this curve may not pass through every bold dots t
almost all those dots lie very closer to the curve.
In this 'chapter we are going to get such a curve known as best fitl
curve to the dots. If we can find such a best fitted curve Y = ¢(x) I
can predict a value of Y corresponding to a value of x. This type .of CUI
is known as best fitted curve of y on x as we considered differerit
corresponding to a single Xi'
CURVE FITTING BY METHOD OF LEAST SQUARE 413
the curve Y = if>(x;a, b,·· ..) is the best fitted curve of the bivariate values
(Xi' Y j} which is known as least square regression curve.
The function if>(x;a,b,··· ..) is called Least Square Regression
Function of Y on X.
The random variable U; = if>(X;a,b,··· ..) is the best representation
of Y by a function of X. The variable Vy = Y - U , is known as Residual
ofY.
Example: In a night school of adult education, let (X, Y) be a bivariate
where X = age of students and Y = No. of days absence.
The average number of days absence is 4 for the students having age
21. That is for X = 21, the mean of the corresponding values of Y be
my (21) = 4. Following is the table of these type of datas :
Age (X) 21 38 42 47 53 61 64
10 14 17 19 34 38
414 ENGINEERING MATHEMATlCS-UA
y = -17.4+ 0.791x
21 38 53 61 64
The co-ordinates of the above dots are (21,4), (38, 10), (42, 14), (47,
17) . ... and so on. Suppose we desire to draw a straight line which will
fit best to the above dots. Consider the family of straight lines y = A + B x
where A, B are arbitrary. It can be proved that among this family of lines
the straight line y = -17 . 41 + 0 . 791 x will be such that the deviation
E[{X-e-17.41+0.791X)}2] will be mimmum. So
y = -17 . 41 + 0 . 791 x will be the best fitted straight line or the least square
regression line of the bivariate values (Xi'Y) and so the function
¢I(x) = -17 ·41 + 0·791 x is least square regression function of Y on X.
Uy=-17·41+0·791X IS the best representation of Y.
Vy =Y-(-17·41+0·791X) is the residual ofY. From this best fitted
straight line y=-17·41+0·791x we can predict the number of days
absence for a student of 50 years as -17.41+0.791x50=22·14:::22
days.
11.3. Normal Equations
For a bivariate (X, Y) if y=¢lex;A,B,C,. .. ·) be a family of curves ;
A, B, C
.
bemg parameters,
as = 0 -as = 0 -as = 0 ..
- are called
aA 'os 'ac
the normal equations where S = E[{Y -¢l(X;A,B,C,' .... rj
415
CURVE FITTING BY METHOD OF LEAST SQUARE
.
equatiOns as
aA -0 as
- 'aB --0 ,etc. then y = ifJ(x; a, b, c,···· ..) is the least
square regression function and so this will give the best fitted curve to
the given data.
Proof. Detail of the proof is not discussed.
Note. In fact the above development is done for finding the least square
regression curve of Y on X.
For least square regression curve of X on Y similar fonnulation do
hold.
Observing the scatter diagram of a bivariate (X, Y) we decide the form
of the curve to be fitted to the datas. It may be straight line or parabola
or any other curve like exponential curve etc. In our next discussion we
shall show how to fit these curves.
11.4. Finding best fitted straight lines.
(i) For a bivariate (X, y), y = a +bx is the best fitted straight line of
_ a a -
Y on X where a = Y - r,y -L and b = rxy -L X
ax ax
\
(ii) x = a + by is the best fitted straight line of X on Y
where a = -X a -
- r ~ _x Y and b
-;a y
= rxv a _x ,where X,Y are mean of
. ay
all values assumed by X and Y respectively.
Proof, Beyond the scope of the book.
The above two formulae can be used for any distribution of (X, lJ·
But if Yassumes unique Yi corresponding to an Xi then the above
formulae can be reduced to more simpler form.
This is shown below.
11.5. Finding best fitted straight line when (X, Y) assumes n pair
of datas (X"y,),(X1'Yl)" .... (xn,Yn)
(i) Regression line of Y on X : When (X, Y) assumes the n pair of
1
values (XI'Yl)'(X2'Y2)····· (xn,Yn) (with same probability -;;) then the
~ 2 2~ 2'
and aB =;~(y; -A-Bx;)(-x;) =-;~(Ax; «s», -x;y;)
.
.. the normal equations -
as = 0 and -as = 0 become
aA aB
L(Y; -A-Bx)=O and L(Ax; + BX;2-x;yJ=O
;
LX; =an+bLY; s
LX;Y; =a LY; +bLy;2
which are known as normal equations for least square regression line.
EM-2~
417
CURVE FITfING BY METHOD OF LEAST SQUARE
Proof. As above.
Note. Making predictions is an important objective of finding least
square regression line. When we have to predict Y for a given value of X,
then we have to use the regression line 'of Y on X in which Y is dependent
variable and depends on the variable X. Similarly, when the problem is of
predicting X, given Y then the regression line of X on Y has to be used.
Example 1. Fit a straight line of the form Y = a + bx to the following
datas:
x 2 5 6 8 9
y 8 14 19 20 31
Predict Y when X = 9 . 6
Solution. Calculation for fitting straight lines
X2 x;Y;
X; Y; I
8 4 16
2
14 25 70
5
19 36 114
6
20 64 160
8
31 81 279
9
:se
~>; =30 I>; =92 LX; = 210 LX;Y; =639
Here n =5.
The normal equations for a and bare
LY; =ax5+bLx;
'of
and LX;Y; =aLx; +bLx;2
Yj
Xi 0 2 3
•
• •
0+1+2+3 6 3
1st Process: For x ==0 , mean of y, my ==--4-- =-=_.
4 2'
0+1+2
For x ==1, m == ==1 ;
y 3
0+1 1
For x==2 m ==--==-
, y 2 2
So we construct the following table.
x o 2
3 1
2 2
CURVE FIITING BY METHOD OF LEAST SQUARE 419
X; m/x;) X2
I
x;m/x;)
• 0
3
-
2
0 0
1 1 1 1
1
2 - 4 1
2
3 3 5 2
Here n = 3.
The normal equations are
Lm/x;)=an+bLx;
and 2 = 3a + 5b i.e. 3a + 5b = 2
•
Solving we get a =~, b =-.!..
2 2
•• the required best fitted straight line is y = ~ - .!.. X
2 2
2nd Process
From a theorem we get the best fitted straight line as
y = a +bx where
- ay - - Cov(x,y) ay-
a=y-rX),-x=y- '-x
ax «», ax
I - Cov(x,y)-
or, a= y-
ax
2 X
1
420 ENGINEERING MATHEMATICS- IIA
Calculation of a and b
2 2
X, Yi Xi Yi XiYi
0 0 0 0 0
0 1 0 1 0
0 2 0 4 0
0 3 0 9 0
1 0 1 0 0
1 1 1 1 1
1 2 1 4 2
2 0 0 0
2 1 ~ 1 2
a} 1 LX2
=- - ( -Lx
1 )2 =-xll-
1 (1)2
-x7 50
=-
n n 9 9 81
- 1 1 7 - 1 1 10
x=-Lx=-x7=-, Y=-Ly=-xl0=-
n 9 9 n 9 9
25
. . 10 81 7 10 1 7 3
.. a=-+-x-=-+-x-=-
9 50 9 9 2 9 2
81
25
. -81 1
and b=--=--
50 2
81
o,
and b =r X)'-
. O"y
25 25
. 7 81 10 7 25 10 9 81 5
.. a=-+-·-=-+-·-=- and b=--=--
9 80 9 9 80 9 8 80 16
81 81
:. the best fitted line of x on y is
9 5
x =a + by or, x = "8-16 y
Proof. Omitted. In fact the proof is similar to that for straight line fitting.
422 ENGINEERiNG MATHEMATICS -llA
x3 X4
2
= Y; s. I I =». 2
x; Y;
0 I 0 0 0 0 b
1 5 1 1 1 5 5
2 10 4 8 16 20 40
3 22 9 27 81 66 198
4 38 16 64 256 152 608
x; y; Y; = log., y; X2
I
x;Y;
l
424
ENGINEERING MATHEMATICS - llA
Example. Fit a geometric curve of the form y = ax" to the following data :
x 1.2345
Y 5·0 6·3 7·2 7·9 8.5
Taking 'log.,' on both side of y = a x" we get
log., y = 10gIO a + b log., x
or, Y = A + b X (1)
where A = loga and X = logx.
We shall find A, b for the best fitted straight line like (1) in X _ Y
frame for the given data
CURVE FITTING BY METHOD OF LEAST SQUARE 425
Calculation of A, b
1 s·o 0·699 0 0 0
2 6·3 0·799 0·301 0·091 0·240
3 7·2 0·857 0·477 0·288 0·409
4 7·9 0·898 0·602 0·362 0·S40
5 8·5 0·930 0·699 0·489 0·650
4·183 2·079 1·17 1·839
LY; =An+bLXi
Solution. The figure against x and yare very large. We shall reduce the
figure in the following table:
(i) Obtain the two least square regression line from the above
information
(ii) A given trip has'to take seven days. How much money should a
salesman be allowed so that he will not run short of money ?
- 1~ 1
Solution. Here X = - L..X = -x 510 = 5
n 102
- 1 1
Y=-LY=-x7140=70
n 102
(i) Regression line ofY on X is
y=a+bx
2
when LY=na+bLX and LXY=aLX +bLX
- a = - 355 = -0.59
601
and b = 48 = 0·080
601
428 ENGlNEERING MATHEMATICS - IlA
x
o 2
x Y X2 l xy
0 0 0 0 0
0 1 0 1 0
.
t ,
0 2 0 4 0
1 0 1 0 0
1 1 1 1 1
2 0 4 0 0
J
CURVE FITTING BY METHOD OF LEAST SQUARE 429
Here n =6
2 1 2 ( 1 )2 1 ( 1 )2 5
ax =-;;Lx - -;;Lx ="6x6- "6x4 "9
t
2 1 2
a=-LY--LY
( 1 )2 1·
=-x6--x4
(1 )2 =q
5
Y n n 6 6
- 1 1 2
X=- Lx=-x4=-
n 6 3
- 1 1 2
y=- Ly=-x4=-
n 6·3
(i) The least square regression line ofY on X is
(J - - Cov(x,y)-
Y =a + b x where a = y - r ---L x = Y 2 X
XY(J (J
x x
x Y X2 x3 X4 xy x2y
0 0 0 0 0 0 0
0 1 0 0 0 0 0
0 2 0 0 0 0 0
1 0 1 1 1 0 0
1- 1 1 1 1 1 1
2 0 4 8 16 0 0
4 4 6 10 18 1 1
an+bLx+c.z>2 = LY
That is, 6a + 4b + 6c = 4
4a + 6b + lOc == 1
6a + lOb + 18c == 1
2
1
or , y == 1--x
2
See the parabola reduces to a straight line.
Example 4. Fit a second degree parabola to the following data taking
x as independent variable, by method of least square:
x o 1 2 3 4
y 1 1· 8 1·3 2·5 6·3
CURVE FITTING BY METHOD OF LEAST SQUARE 431
Find out the difference between the actual value of y and the value of
y obtained from the fitted curve when x = 2
x y x-3 y-l·3 10(y-l·3) u2 £13 u4 £IV u2v
=£1 =v
0 1 -3 -0·3 -3 9 -27 81 9 -27
1 1· 8 -2 0·5 5 4 -8 16 -10 20
2 1· 3 -1 0 0 1 -1 1 0 0
3 2·5 0 1· 2 12 0 0 0 0 0
4 6·3 1 5 50 1 1 1 50 50
2
otal -5 2>=64 15 -35 99 49 Lu v
2 3 4
= LU F Lu Lu = Lu = LU~ =43
an+bLu+cLU2 = Lv
aLu+bLu
2
+cLu
3
= LUV
3 4
aLu2 + bLu + cLu = Lu2v
That is, Sa - 5b + 15c = 64
-Sa + 15b -35c = 49
15a - 35b + 99c = 43
Solving these three equations we get
Exercise 11
1. Fit a straight line to the following data by method of least square:
x o 1 2 3 4
,y 1 1·8 3·3 -1·5 6·3
2. The following table gives the age of cars of a certain make and
annual maintenance costs. Find the least square regression line for costs
related to age:
Age of Cars 2 4 6 8
(in years)
Maintenance 10 20 25 30
cost (in hundreds of Rs.)
3. Find the equation of the best fitted straight line for the following
bi-variate data :
x 10 12 13 16 17 20 25
y 19 22 24 27 29 33 37
4. Find the equation of the best fitted straight line for the following
observations :
x 4 6 8 12 15 17 22
y 46 42 40 36 30 25 19
S. Find the least square regression line for the following data
(i) x 70 80 90 100 118 120
y 553 547 539 533 527 520
CURVE FITTING BY METHOD OF LEAST SQUARE 433
(ii) X 5 7 9 11 13 15
X 2 4 6 8 10 12 14
Y 10 14 15 16 15 17 18
7. The following data relate to results of a fertiliser experiment on crops
yields:
Units offertiliser 0 2 4 6 8 10
used (x)
Units of field (y) 110 113 118 119 120 118
Fit a straight line to the above data and estimate the amounts of yeild
when units of fertilisers used are 3 and 7 respectively.
EM-2A-28
434 ENGINEERING MATHEMATICS -IIA
16. At the ends of years 2005 through 2012, a company had the
following investments in plants and equipment :
1·0, 1·7, 2·3, 3 ·1, 3·5, 3·4, 3·9 and 4·7 million dollars.
Fit a least square line. Assuming that the trend continues, predict the
company's investment in plant and equipment at the end of 2014.
17. Find both the least square regression line from the following:
Ly 2
= 1720, and n = 10
18.From the following results, obtain the two least square regression
line and estimate the yield, when the reainfall is 29cms and the rainfall,
when the yield is 600 kg.
21. By least square method fit a parabola to the following bivariate data
(i) x 1 2 3 4 5
y 11 30 59 98 147
(ii) x -1 0 1 2 ·1 3 4·2
Y 3 -1 ·9 10·13 23·1 47·7
(Hi) x 0 2 4 6 8 10
y 2 ·1 -2·1 2 ·1 14 ·1 34 ·1 62 ·1
(iv) x -1 0 1 2 3 4
y 4 2 6 16 32 54
22. The profit y ( in lakh of rupees) of a certain manufacturing company
in the year x are given in the following table:
x 2 3 4 5
(i) x -1 0 3 5 7
y 1· 5 3 24 96 384
(ii) x -1 -2 0 1 2 3 4
Y 8 16 4 2 1 0·5 0·25
(iii) x 1 1· 5 2 2·5 3
y ·6 ·848 1·2 1· 70 2·4
(iv) x -1 0 2 4 5 6
y 1· 5 3 12 48 96 192
(v) x 2 3 4 5 6
27. Fit the geometric curve of the form y=axb to the following
bivariate datas :
(i) x 1 2 3 4 5
y 3 48 243 768 1875
(ii) x 0 1 4 9 16 25
Y 0 ·5 1 1·5 2 2·5
(iii) x 4 5 6 7 8
y 8 12·5 18 24·5 32
(iv) x 1 2 3 4 5
Y ·2 1· 6 5·4 12·8 25
,
I
-J
438 ENGINEERING MATHEMATICS - I1A
Answers
1. y = 11. 96 - 4 . 29 x
2. Y = 3 . 25 x + 5
3. y=7·75+1·21x
4. Y = 52 -1· 5x
6. y=1O·72+0·536x
7 17 153( 7)
,20. y-9= 60(x-4);x-4= 50 y-9
439
CURVE FITTING BY METHOD OF LEAST SQUARE
2
21. (i) y=2.-t4x+5x2 (ii) y=-I.02-I.Ox+3.0x
2
(iii) y = 2·04 - 4· OOx-1· Ox2 (iv) y = 2·0 + 1· Ox ., 3x
22. Y = 2 ·048 + ·08Ix + ·055x2
23. y = 1140 + 72 -Ix + 32· 2X2; Rs 2732
2
24. Y = 9 . 8 - 2 . 60x + 0 . 208x
n ;=1
is called a Sample Variance.
The positive square root, i.e. S is called a Sample Standard Deviation.
Population Mean and Population Variance. The A.M of all the observations
in a population is called Population Mean.It is denoted by )..l •
The variance of all the observations in a population is called Population
Variance.!t is denoted by 0"2 •
The positive square root of population variance is called Population S.d.lt
is denoted by 0".
S= 14 = 2 is sample S.d.
Again, )..l = ..!..(2 + 6 + 10 + 14) = 8 is the population mean;
0"2
1 4
="4(22 +62 +lO2 +142)-
{I"4(2+6+10+14) }2 =84-64=20 IS
population variance.
is called a statistic.
For example X[ + xi +...+x~ is a statistic.
Illustration. Let {2,4,5} be a population. Consider the samples {Xj,X2}
J
444 ENGINEERING MATHEMATICS - IIA
. _ 7 9 7 9
1. e. 3 - 3 - - -
X •
. '2' '2' 2 '2
Similarly,sample variance, sample standard deviations etc. are all statistic.
var(x)=E(X2)_{E(X)}2=2~2-82=230:. o-x=J¥=2.58.
(J
(ii) f..l and r if N is infinite
. vn
(J
(iii) f..l and .[;z if the sampling is SRSWR.
Note. (1) The mean and S.d of the approximate normal distribution is
according to Theorem-2.
11 __ 11+8+6+3+2 --6.
The mean of the population, ,.-
5
The variance of the population
Probability:
1
2 5 ~ -2 -
I
-4 2
- -
I
-2 2
- -
2 2
----
I 2 I
25 25 25 25 25 25 25 25 25 25 25 25
The mean"of this sampling distribution
1 2 2 1" 4 2
= 11 x-+9.5x-+8.5x-+8 x-+7 x-+6.5x-
25 25 25 25 25 25
1 2 2 2 2 1
+6x-+5.5x-+5 x-+4.5x -+4 x -+3 x-
25 25 25 25 25 25
+2.5 x - 2 + 2 x - 1 = 6 = Ponulati
opu anon mean.
25 25
/
/
SAMPLING & ITS DISTRIBUTION 449
from a population with mean ~ and S.d 0- then E( 82) = n- 10-2 (where
n
population size is infinite or the sample is drawn with replacement).
;=1
2
... (1)
S'- = -1 ~(
L.J X; - ~ + ~ - X_)2
n ;=1
EM-2A-29
450 ENGlNEERlNG MATHEMATICS -IIA
1 II cr2
=-}2Var(x;)--
n ;=1 n
1 II cr2 1 cr2
= - 'Lcr2 -- = -·ncr2--
n ;=1 n n n
cr2 n-l
:::;cr2--=--cr2
n n
Note. Though in the above theorem we found the mean of the sample
variance S2 we did not get the sampling distribution of the statistic S2.IL
the following theorem we go for those.In fact we get the sampling distributior
nS2
of which is sufficient to characterise S 2.
-2-
c
Theorem 2. If S2 be the sample variance of a sample of size n draw
nSl
from a normal population with mean Il and S.d o then the statistic -,cr-
size 10
(i) taken with replacement
(ii) taken without replacement.
Here sample size n = 10 ,population size N = 112 ,populatiOlyS.d,
c = 13.2 .The distribution of the population is not known.
·) Sd f-- o _13.2_13.2_417
CI . 0 x- ..r;;- .JiQ-'3.16- .
Example 4. Compute the standard error of the mean and construct the
sampling distribution ofthe mean for simple random samples of two families
each from a population of 5 families which is given below:
Family A B C D E
Family size: 4 3 2 5 7
below:
Sample (4,3) (4, 2) (4, 5) (4, 7) (3, 2)(3, 5)(3, 7) (2, 5) (2, 7) (5, 7)
4 5 3.5 4.5 6
3.5 3 5.5 2.5
SAMPLING & ITS DISTRIBUTION 453
Since the population is normal and the sampling is random so the sampling
distribution of the statistic x (sample mean) is normal distribution with
mean = population mean = 0.1 and standard deviation
So, z =
r -u:
----;yf- is
--~+---~----~~Z
standard normal variate. 1.43
From table (shown at the end of the book) the area under the normal
curve enclosed between y-axis and the ordinate z = 1.4 is .4192.
So the area under the normal curve enclosed between y axis and the
ordinate z = -1.4, = 0.4192 .
So the required probability = OJ - 0.4192 = 0.808.
SAMPLING & lTS DISTRIBUTION 455
So, P(X>975)=P(Z>-~)
-..In
5
456 ENGINEERI G MATHEMATICS -IIA
- 68.3-68
When X=68.3, Z= =·5
.o
:. Probability of X lying between 66.8 and 68.3
y
z
0.5
= P(66.8 < X < 68.3)= P(-2 < z<.5) = P(-2 < Z < O)+p(O<z <.5)
= 0.4772 + 0.1915 = 0.6687 .
So,the expected number of required type of samples
= 80 x 0.6687 = 53.496::::53 .
SAMPLING & ITS DISTRIBUTION 457
Example 9. The mean weight of 500 ball bearings is 5.02 gms.Their S.d is
0.30 gm.Find the probability that a random sample of 100 ball bearing
2.96
would have a combined weight more than 510 gm.[given f <j>(z)dz = 0.4985
where Hz) is standard normal function] 0
The population is not given as normally distributed. The sample size 100
is larger (> 30) . So by Central limit theorem the statistic X is approximately
normally distributed with mean == populati<?nmean = 5.02 and S.d
When
- z- 51-502=2""
X = 5.1, 0017 :>u.
~
Probability that the sample has combined wt. greater than
510= p(X > 5.1) = Piz > 2.96)
= area under standard"normal curve enclosed on right side of the ordinate
z=2.96 (shown in the figure by shade) =0.5-0.4985=0.0~15.
2
2 (61- Il) 2 (80 - Il)
When x = 61, Z =
(j
and when x = 80, Z =
(j
.
2(61-1l))
.', From (1) we have, P ( Z < o = 0.44
"
2(1l-61)) (3)
or, P(0 <Z < • o =.06
_ 26 (8)2 _ 26 64 _ 14
--- - ------
3 3 3 9 9
Then ~O'2 = 2-1 x 14 =2
'n 2 9 9 (2)
3 2 1 2 9 2}2
{ Ox-+lx-+-x_+_x_
9 9 4 9 4 9
= {3.+~ 3.+~}_{3.+~+J..}2
9 16
x
9 16 9 18 2
49 49 245
=---=_.
36 81 324
:. s
O' ! =J:;: =0.87.
460
ENGINEERING MATHEMATICS-IIA
Example 12. Samples of weights of 200 students each are drawn from a
very normal population of the weights of students with S.d 10 pound.
Variance in each sample are computed.Find (i) the mean and (ii) the standard
deviation of the sampling distribution of sample variance.
If S2 is sample variance then.
The required mean
.'. vaf~~n=2X(200-')
or
,
Var(S2)= 2x 199
4
= 199
2
:. c
s
2 =t 99
2'
Example 13. From a normal population of S.d ..fiO samples of size 5 are
drawn. Prove that the probability that the sample variance are greater than
2.336 is O.90.Find the mean and variance of the sampling distribution of
the sample variance. [Given X~.90.4 = 0.584 ]
Here 11 = 5, c =..fiO .Let S2 be the sample variance. Since the population
S2
is normally distributed so n 2 has X2 distribution with 5 -1 =4 degrees of
c
,
freedom.Given X~.90.4 = 0.584 . This implies that for a X- variate with 4 degrees
or, {~)' > 058~) = 090 or, p( ~' > 0584) = 090
or, p(S2 ~ 2.336) = 0.90 .Hence proved.
SAMPLING & ITS DISTRIBUTION 461
. 2 I1S2 5S2 S2
Since the X variate -2 = ---2 =- has 4 degrees of freedom so
c (Fa) 4
or, p( lOt > 3.325) ~ 0.950 or, p( S' >1.995) ~ 0.950 (1)
1.995 10.146
462
ENGINEERING MATHEMATICS - JJA
Exercises 12
1. Ages of 5 persons have been recorded (in years) as 14, 19, 17,
20, 25.For random samples of size 3 drawn without replacement from
this population obtain the sampling distribution of x. Show that the
mean of x equals the population mean and obtain the S.E of x, directly
from the sampling distribution and also by using the formula.
[Hint : the rrean of the distributionof x is 19 = populationmean]
2. The values of the characteristic x of a population containing 6 units
are given by 2,6,5, 1, 7, 3.Take all possible samples of size two and verify
that the mean of the population is exactly equal to the mean of the
statistic,sample mean.
[Hint: Mean of x = 4 = population mean]
3. A simple random sample of size 4 is drawn with replacement from a
population with S.d 6.What is the standard error of the statistic sample
mean?
J
SAMPLING & ITS DISTRIBUTION 463
15. The radii of the ball bearing is normally distributed with standard
deviation JlO .Samples of size 15 are drawn.Find the mean and S.d of the
sampling distribution of the sample variance.Find the probability that the
sample variance does not lie between 2.90 and 4.70.
of the sample variance are respectively ~ and 275 .Find the variance of
12 72
the population and the size of the samples drawn. Then fmd the probability
of sample variance exceeding l3.025. [Given that X]JOl;11 =31.26]
Answers
5J5
3. 3 4.4, "9' 3 5.0.2164
13. (i) 0.5 (ii) .0228 (iii) 0.6~72 14. ~ , Jl9; .02
28 112 8
15.3" 9; .955 16. 5, 12; .001.
SAMPLING & ITS DISTRIBUTION 465
EM-2A-30
-
466 ENGINEERING MATIIEMATICS-IJA
(a) the size of the population goes on decreasing after each drawing.
(b) the size of the population does not go on decreasing after each
drawing.
(c) the sample size becomes equal to that of the population size.
(d) none of these.
10. An observation may repeat in a
(a) SRSWR sample (b) SRS WOR sample
(c) any type of sample (d) none of these.
11. Under SRSWR the possible samples from the population {2,6,7}
are
(a) {2,2},{2,6},{6,7}
(b) {2,6},{2,7},{6,7}
(c) {2,2},{6,6},{7,7},{2,6},{2,7},{6,7}
(d) none of these.
SAMPLI G & ITS DISTRIBUTION 467
12. Under SRSWOR all the possible samples from the population
{x,y,z} are
5 3
(a) 3 (b) - (c) - (d) 2.
2 2
17. Sample standard deviation is
18. If {Xi ,x} } be sample drawn from the population {Xl ,X2'X3' ··,xIO}
{i,j = 1,2,3, ...io} then which one of the following is not statistic
2 2
(b) Xi +Xj
(d) X2
I +x/. _
468 ENGINEERING MATHEMATICS-IIA
19. If {x;,xj Hi,} = 1,2,3,.· ·10} be sample drawn from the population
{x ,X2,X3""'XIO},
j then which one of the following is not parameter.
(a) x; + x j (b) x/ + x /
9 10
(c) s, + LX; (d) LX; .
;;1 ;;1
20. If {6} is a sample drawn from the population {2,4,6,8} then sample
standard deviation is
(a) 1 (b) 0.5 (c) 0 (d) 0.8.
21. Which one of the following is incorrect
(a) s.d of a sample is a statistic.
(b) mean of a sample is a parameter.
(c) variance of population is a parameter.
(d) s.d of population is a parameter.
. 22. Sampling distribution is the
(a) distribution of population.
(b) distribution of a sample.
(c) distribution of a sample statistic.
(d) distribution of a parameter.
23. If a sample of size two is drawn from a popuation {4,6,8} at a
time then the sampling distribution of the sample mean is
(a) X 10 12 14
1 1 1
f - - -
3 3 3
1 1 1
(b) X - - -
5 6 7
2 1 1
f .' - - -
3 3 3
(c) X 5 6 7
1 1 1
J; - - -
3 3 3
(d) none of these.
SAMPLING· & ITS DISTRIBUTION 469
33. The s.d of the sampling distribution of sample mean for the population
with s.d 0.5, population size 122 and sample size 22 is
:; 05
(a) 22 (b) ~
5
(c) 11m (d) none of these.
34. For any large population, for large values of sample size the sampling
distribution of sample mean has
(a) exactly normal distribution
(b) approximately nonnal distribution
(c) uniform distribution
(d) t-distribution.
35. The sampling distribution of the sample mean for a large population
is approximately normal if the sample size is
(a) 2 (b) 5
(c) 10 (d) 100.
36. Consider all possible samples of size two that can be drawn with
replacement from a population with mean 6 and variance 10.8. Then the
standard error of the sample mean is
(a) 9 / (b) 8
(a) 40 (b) 45
43. The possible number of samples of size two drawn with replacement
from a population of size 25 is
44. A normal population has a mean 0.1 and s.d 2.1. The mean of the
sampling distribution of the sample mean with sample size 900 is
45. A normal population has s.d 2.1. The standard error of the sampling
distribution of the sample mean with sample size 900 is
normal variate
(a) True (b) False.
47. The mean weight of 500 guinepigs is 5.02 gms and their s.d is 0.30.
Samples of 100 guinepigs are drawn. The s.d of the sample mean is
(a) 0.27 (b) 27
197 199
(c) - (d) -
2 2
Answers
l.c 2.a 3.b 4. a 5.d 6.b 7.b 8.c
Then we see under this hypothesis the entire character of the population
is not known because the standard deviation remains unknown. So this
hypothesis " J.l = 60" is a composite hypothesis.
Null Hypothesis
Test of Hypothesis starts with a statistical Hypothesis. A statistical
hypothesis whose possible acceptance or rejection is tested on the basis
of sample observation is called a Null Hypothesis. Usually it is denoted by
Ho·
Illustration. Suppose we assume that "the mean of a population is 40".
Let a random sample drawn from this population has mean 38. By going
through some statistical analysis we have to test whether our assumption
may be accepted or rejected. Then the assumption" J.l = 40" is a Null
Hypothesis; in notation Ho(J.l = 40).
Alternative Hypothesis.
A statistical hypothesis which is different from the Null Hypothesis is
called Alternative Hypothesis. It is denoted by HI .
Illustration. Consider the null hypothesis Ho (J.l = 60) . Then HI (J.l :j; 60)
is alternative hypothesis. Again if it is seen from the fact that there is no
chance of J.l being less than 60 then we may also take HI(J.l > 60) as
an Alternative Hypothesis to Ho(J.l = 60). Similarly HI(J.l < 60) may be
also considered as Alternative Hypothesis.
Both Sided Alternative Hypothesis.
Let Ho (8 = 80) be a null hypothesis where 8 is a parameter. Then
the alternative hypothesis HI (8:j; 80) is called both sided alternative
hypothesis.
For example, the alternative hypothesis HI(J.l:j; 60) is both sided
alternative hypothesis against the null hypothesis Ho(J.l = 60), where J.l
is the mean of the population.
One Sided Alternative Hypothesis.
Let Ho (8 = 80) be a null hypothesis where €I is a parameter. Then
the alternative hypothesis HI (8 > 80) is called one sided (Right) Alternative
'Hypothesis. The alternative hypothesis HI (8 < 80) is called one sided (Left)
Alternative Hypothesis.
TESTING OF SIGNIFIANCE 475
For example the alternative hypothesis HI(~ > 60) is one sided
alternative hypothesis against the null hypothesis Ho(~ = 60) .
13.3. Test Statistic.
To test the possible acceptance or rejection of a null hypothesis we
have to take the help of a statistic whose sampling distribution is known.
By evaluating the value of the statistic we decide whether the ulI
Hypothesis is to be rejected or not. This statistic is called Test-Statistic.
Illustration. Let Ho(~ = 60) be a Null Hypothesis where Il is the
population mean. Let the population be normally distributed with s.d.
• c = 3. We consider the statistic x,
the sample mean. Its sampling
3
distribution is known to be normal with mean 60 and s.d. where.J15
15 is the size of the sample. We compute the value of x
for a sample
drawn from this population and from this computed value we decide
whether Ho is to be accepted or not. So here x is the test statistic.
Note: Selection of test statistic depends on the character of population,
sample and the Null Hypothesis.
13.4. Critical Region, Region of Acceptance and Level of
Significance.
Let Ho be a null hypothesis and t be the appropriate test statistic by
which we decide whether Ho would be accepted or not. This decision is
based on probability consideration. Usually a lower probability a. (e.g 0.05
or 0.01 etc) is taken under consideration and let P(a ~ t ~ b) = 1- a.. If
a computed value of t lies within this interval (a,b) then 'we decide Ho
is accepted at "a. Level of Significance" or " lOOo.% Level of
Significance". The interval (a,b) is called Region of Acceptance
corresponding to" a. Level of Significance".
If a computed value of t lies beyond the interval (a,b) we decide Ho
is rejected at " a. Level of Significance" or " 100 a. % Level of
Significance". This region beyond the interval (a,b) is called "Critical
Region" (CR) of the test at" a. Level of Significance".
Note: (1) The region of acceptance/Critical region is not unique
corresponding to a 'Level of Significance'.
476 ENGINEERING MATIIEMATICS-IIA
p( -1.96 < z < 1.96) = 0.95 = 1- 0.05 (obtained from statistical table) so it
is best to consider (-1.96,1.96) as the region of acceptance at 5% level
of significance and the region beyond it. i.e. the region Izl ~
1.96 as the
critical region at 5% level of significance. (The reason of such selection
would be discussed later)
Now if it is seen that a random sample (of size 400) gives its mean,
. 171.38-171.17
x = 171.38 then we compute the value of z = = 1.27 .
0.165
TESTING OF SIGNIFIANCE 477
conditional probability)
= Level of Significance of the test.
This is also called Size of Type Error
478 ENGINEERING MATHEMATlCS-IlA
L/
TESTING OF SIGNIFIANCE 479
To obtain the best test we find a critical region. (e.g the region where
It I "> a) corresponding to a level of significance a such that Probability
of Type I error relative to this CR is a i.e. P (computed value of t lies in
the C.R assuming 8 = 80) = a. There may exist several CR like this.
) Among these that critical region is called Best Critical Region for which
(
the Type II error is least. We say this is the Best Critical Region or simply
Critical Region (CR) corresponding to a level of significance.
480 ENGINEERING MATHEMATICS-IIA
which is standard normal variate. Take the level of significance of the test
as a=O.OI.Assuming J.l=52 we see z=x-52.
From statistical table we have
p(O < z < .03) = 0.01
p( -.900 < z < -.806) =.3159-.3051 =.01
Power of a Test ..
From the previous discussion we see that the goodness of a test depends
upon the choice of the critical region. As the probability of Type II error
decreases the goodness of the test increases. This notion leads to have
the definition:
So if the Best Critical Region can be selected the Power of the Test
becomes highest.
Illustrative Example.
Example 1. In order to test whether a coin is perfect the coin is tossed 5
times. The null hypothesis of perfectness is rejected if more than 4 heads
are obtained What is the probability of Type I Error? Find the probability of
Type II Error when the corresponding probability of head is 0.2.
[ W.B. U. Tech 2007]
Let P be the proportion of the number of obtained head among 5 times
of throw. The coin is perfect if P = ± since probability of" head for a
. =SCS(~)S
2
(1- ~)S-S =_1
2 32 .
Now, Probability of Typer II Error when P = 0.2
= Probability of acceptance of Ho assuming P = 0.2
= P(j s 4) assuming P = 02 = 1- P(j = S) assuming P = 02
= l-sCs(o.2)5(1- 0.2)5-5 = 0.99968 .
Ei\J-2A-31
,...
I
482 ENGINEERING MATHEMATICS -IIA
Exercises 13
1. (a) In order to test whether a coin is perfect, I shall toss it 6 times.
I shall reject the null hypothesis of perfectness if and only if I get no head
or 6 heads. What is the probability of Type I error for my test?
(b) To test the unbiasedness of a die it is thrown six times and is
accepted that the unbiasedness if not more than one sixes are obtained. Find
the probability of Type I Error. [ W.B.UTech 2005 ]
= 0 , elsewhere
and that you are testing the null hypothesis Ho:9 = 1 against HI:9 = 2 by
means of a single observed value XI. Determine the size (i.e. probability)
of Type I and Type II error if you chose the interval 0.5:s; x as the critical
region.
Answer
1. (a) 1/32 (b) 0·263 3. 0.00008448; 0.99870967
11
4. 64,1-0.382=0.6185. 0.5,0.25
TESTlNG OF SIGNIFIANCE 483
(a) H(~ * 5) (b) H(~ < 5) (c) H(~ > 5) (d) H(~ = 4)
8. If ~ is a parameter and H(~ = 7) is null hypothesis, then which one
of the following is Both sided alternaive hypothesis
15. 'Falling of the computed value of the test statistic critical region' =
complement of the even 'Falling of the computed value of the test statistic
in the region of acceptance.
16. If in a test of hypothesis (-00,1.02) is critical region at .0 I level of
significance then, if t is test statistic,
(a) p(O < t < 1.02) = 0.01 (b) P(1.02 < t) = 0.99
(c) p(O < t) = 0.96 (d) none of these.
17. If the alternative hypothesis is taken right sided with test statistic t
then indicate which one of the followings may be a possible region of
acceptance
Answers
29.b
r
111411;;;;;;;;;;;;;;;;;;;;;L;;;AR;;;;;;;;;G;;;E;;;S;;;A;;;MP;;;;;;;;;L;;;E;;;T;;;E;;;S;;;T;;;O;;;F;;;;;;;;;SI;;;G;;;N;;;IF;;;I;;;C;;;AN;;;C;;;;;;;E
14.1. Introduction:
The sampling distribution of many of the commonly used statistic is
almost normal when the statistic is measured on a large sample
(approximately more than 30). For example for a large sample the statistic
Theorem 1. Let a large random sample XI ,X2 ""Xn (n > 30) be taken
from a normal population with mean J.1 and s.d. cr to test the Null
Hypothesis Ho(J.1 = J.1o) against y
an alternative hypothesis HI at
a -level of significance, where
o is known (i.e. the hypothesis
are simple).
o
Then the Best Critical Region (CR) determined by the test statistic x (the
sample mean) is
. X-J.1o
i.e., Z = cr/.J;, < -Za if HI is Left sided.
.
i.e, Z = cr/.J;,J.1o
X - > Za 1if H I IS
' Righ t SIide d .
LARGE SAMPLE TEST OF SIGNIFICANCE 489
(iii) Ix - 1101> J-;; Za/2 i.e. Izi = :i?n > Za/2 if H is both sided,
J
s.d. 0.1.
Note that this is a Right Sided hypothesis. The test statistic is x
(sample mean) which is normally distributed (as discussed in an earlier
(J 01 x-2
chapter) with mean 11= 2, s.d = r - ~ =;()1 :.Z
is standard = --
,,1m "n .01
normal variate. Let the testing is to be done at 0.01 level of significance.
As we see from the above theorem the Best Critical Region is Z > 2.32
at .01 level of significance since .01 is the area under the standard normal
curve enclosed between the ordinate Z = 2.32 and Z = 00 as shown in the
adjacent figure. This is found from Table I given in Appendix
If for a random sample, its mean x is 2.015 then the computed value
-1.96 1.96
1,;(1' -100)
•• C.R is > 1.64
10
1,;(1' -100)
•. the region of acceptance is < 1.64 .
10
.JnCOU + IOS-100)
Putting this in (1) we get z = _-'---'.In_n ---'-
10
.In( lO"-n +
"tin
U s) IOU +S.Jn
=
10 10
:. From (2) Probability of Type II Error
IOU +S.Jn )
By problem, P ( 10 < 1.64 = 0.02
164-S.Jn )
or, P ( 10 < U < 0 = 05 - 0.02 =.48
s.d.
=~ 100 = 14 14
.r;; = J50 .. So,
z_
-
x -14.14
1800 .
IS standard normal vanate.
.
Since HI is right sided, the best critical region at 0.01 level is z » 2.33
(from the supplied data). For the sample the computed value of
1850-1800
z = 14.14 = 3.54 which lies in the CR. Thus Ho is rejected and so
Since we are given p(lzi > 2.58) = ·01 so the Best CR is Izl > 2·58
x-120
:. z = --2- is standard normal variate. Since HI is left sided, the
Best Critical Region is z < -1.645 at 5% level of significance
(.,' from statistical table we see area under the standard normal curve
enclosed between the ordinates z = 0 and Z= 1.645 is 0.45)
105-120
The computed value of z = 2 = -7·5 which lies in the CR.
( ...) I I =
111 Z ~p - Po > Za/2 1'fH I IS
. b oth SIidedwere
h Q0 = 1- D
'0 an d a IS
.
PoQo
n
the area under standard normal
curve enclosed between the
ordinates Z = za and Z = 00
as shown in the adjacent figure
Proof . Omitted.
Illustrative Example :
Example. 1. In a random sample of size 400 there are 80 defective items.
Test at 5% level whether the proportion of defective items in the population
1
may be regarded as "6
1·96
[Given f<l>(t)dt=0.475,<I> is the pdfofnonnal variate]
.) is H{P;C~)
The obtained sample proportion p = 80 = O· 2
400
496 ENGINEERING MATHEMATlCS-IIA
n
=
~
6 = p - o· 167
0·186
400
This is both sided test.
. From the given data we have P(O < z < 1·96) = ·475
or, P(I.96< z < 00) = ·5--475 = ·025
By symmetry P( -00 < z < -1· 96) = ·025 also
:. P(lzJ > 1· 96) = 2 x ·025 = ·05
:. The Best Critical Region is JzJ> 1·96
.
Here the computed value of z IS z = 0·2 -0·167
= 1. 77 which is not
0·186
greater than 1.96.
:. z does not lie in the cirtical Region. So H 0 is accepted at 5% level.
• 1
We conclude 'proportion of defective items in the population is (5 ,
Example. 2. A sample of size 600 persons selected at random from a large
city shows that percentage of male in the sample is 53%. It is believed that
male to total proportion ratio in the city is
continued by the observatiors.j,
+. Test whether this belief is
-1·96
We are given' f <1>( z) dz = ·025
2·33
We are given J <I>(z}1z = O·99, 2·33
EM-2A-32
498 ENGINEERING MATHEMATICS -IIA
Example. 5. A die was thrown 9000 times and of these 3220 yielded a 3 or
4. Is this consistent with the hypothesis that the die was unbiased?
2·58
[Given J 4> = 0·95 ]
-2-58
Solution : If the die is unbiased then the probability of" 3 or 4" is
hypothesis HI ( P :;t: ~ ) • This is both sided test. The sample size n = 9000 ;
n
~::CXi_x)2 nS2
(ii) X2 =..:....i=....:..I _ (S is sample s.d) is given by X2 > X~ ,
O"~ O"~
.. X2 = -21 ~( )'
Th e test stanstic L... Xi - J1. -
0" i=1
--
LARGE SAMPLE TEST OF SIGNIFICANCE 501
and {x;,X2, ...x~J, be drawn from two Normal populations (~hcrl) and
(~2 , c 2) respectively to test the hypothesis Ho (~I = ~2) at a level of
significance ;
XI -X2
Then z= is the test statistic which is standard normal
0"2 O"Z
variate. _I +_2 z-curve
• nl n2
I I
(iii) z > za/2 if HI (Ill ':F- 1l2) (both sided)
where a. is the area under standard normal curve enclosed between the
'ordinates Z = za and z = 00 as shown in the adjacent figure
Note. In the above theorem the two population mayor may not be
Normal.Moreover if the population s.d o , and O"z are not known then
we may think 0"1 ~S1 ,0"2 ==S2 where SI'SZ are sample standard
deviations.
14.6. Illustrative Example's.
Example. 1. The means of two large samples of sizes J 000 and 2000 are
67.5 and 68.0 respectively. Test the equality of means of the two normal
populations each with s.d. 2.5 at 0.5% level. Given area under st. normal
curve is 0.25 enclosed between the ordinates z = 0 and z = 0.0987.
Let the two population means be III and 112'
Since HI is taken, both sided so from given data the critical region is
Iz I>.0987 at 5% level of significance. Here for the computed value
Iz 1=1-5.16421>.0987 . So Ho is rejected. We conclude the two population
means are not same.
Example. 2. A college conducts both day and night classes intended to
be identical. A sample of 100 days student yields examination results as
under.
x = 72.4; ax = 14.8
A sample of 200 night students
yields examination results as under:
x=73.9; ax =17.9
.0987
Are the two means statistically equal at 10% level 1".0981
Given p{ 0 < z < l.645)=.05.
Take the null hypothesis HO{1l1 = 112) against HI{1l1 ~ 112) where 111>112
are the two population means.
Here, means of the two sample, XI = 72.4, X2 = 73.9
s.d. of the two samples, SI = 14.8, S2 = 17.9
size of the two samples, nl = 100, nz = 200 .
Since the sample size are too large so though the s.d of population are
unknown we take the test statistic z = XI - X2 which is standard normal
S2 S2
variate. _I+~
nl nz
Computed value of z = 72.4-:-73.9 = - 0.77 .
{14.8)2 (17.9)2
--+--
100 200
Since HI is both sided and since p{O < z < 1.645)= 0.05 so the critical region
is Iz I> l.645 at 10% level. Since 1-0.771= 0.77 does not lie in this region,
Ho is accepted at 10% level.
14.7. Test for Difference of Proportions (i.e. Test of equality of
proportion)
Equality of proportions of a particular type of items in two populations
are tested by drawing two random samples from the two respective
populations.
504 ENGINEERING MATHEMATICS - JIA
l 2 n n :. Xl + X2
Then Z = -r==========, where P = .is the test
1 1) nl +n2
p(l- p) (-+-
n1 n2
statistic which is standard normal variate.
Accordingly the CR. is
(i) Z < -za if HI(~ < P2) left sided
(ii) Z > za if HI(~ > ~) right sided
(iii) Izi > za/2 if HI(~ = ~) both sided
00
~_X2
P(l-P)(~+~) ,
nl n2
A XI +X2 .. .
where P = -+-, which IS a standard normal vanate.
nl n2
~(1-~)(6~0 + 10~0)
So, the computed value 2.66 does lie in this CR .We do not accept
Ho at 5% level and HI' is accepted. We conclude PI "# P2 i.e. the two
cities have not same percentage of literacy.
Example. 2. A company has the head office at Kolkata and a branch at
Bombay. The personnel director wanted to know if the workers at two
places would like the introduction of a new plan of work and a survey
I
r.
was conducted for this purpose. Out of a sample of 500 workers at Kolkata
62% favoured the new plan. At Bombay out of a sample of 40q workers
41 % were against the new plan. Is there any significant difference between
i the two groups in their attitude towards the new plan at 5% level. Given
area under standard normal curve enclosed between the ordinates z = 0
and z = l.96 is 0.475.
Let lj = Population proportion of workers who favoured at Kolkata.
P2 = Population proportion of workers who favoured at Bombay Office.
•• "there is a significance difference between the two groups in their
attitudes towards the new plan" ~ , lj "# Pz '
We shall test the null hypothesis Ho(lj = Pz) against HI(lj "# Pz).
506
ENGINEERING MATHEMATICS _JlA
z== nz
nJ ~ xJ+x2
where p==~
Ip(J- p)(~+~)
V, nJ n2
nl +».
which is a standard nonnal variate.
Corresponding to the two samples
~ 310+236 546 91
p==500+400 ==900==150 and the computed value of
310 236
---
z :=: 500 400 == 3 ==~ ==0.915
J 91 (1 91)( 1 1)
150 - ISO 500+400
.0328x 100 3.28
.
of product IOn. .
Test at 1% level. GIven .[2; !
1 2.33 -Z2/2
e dz==0.49
Since HI (~ > P2) is right sided. From the given data we have
P(z> 2.33) :=:0.5-0.49=0.01. TheCRis z>233.
LARGE SAMPLE TEST OF SIGNIFICANCE 507
1 1 1
I 20
2184
--x-
1
15625 576
~
375000
~~xp;~o ~~xJ~1.21~3.21
I
Let ~,P2 is the two proportions.
1
( We take null hypothesis H 0 (lj = P2)
Alternative hypothesis is HI (lj> P2) .
SI -S2
The test statistic, Z = --;=========-
S2 S2
_I +_2
2nl 2n2
= ---;=2=5
=-=27= = -2 = -0.452
252 272 .J19.529
-+-
60 80
Since HI is left sided so the best critical region (CR) is z > -z.05
8. The mean yield of wheat from district A was 210 1bs. with s.d = 10
lbs. from a sample of 100 plots. In another district B, the mean yield
was 220 lbs with s.d. = 12 lbs from a sample of 150 plots. Assuming
that the s.d of yield in the entire state was 11 lbs, test whether there is
any significant difference between the mean yield of crops in the two
districts. Given pO
z / > 258) =.01 .
[Hint: Here. population (J is known. Take HO(J.l1 = J.l2) against
HI(J.l1 *' J.l2) ]
9.Intelligence test on two groups of boys and girls gave the following
results:
Mean S.D. Number
Boys 70 20 250
Girls 75 15 150
Is there a significant difference in the mean scores obtained by boys and
girls? [ W.B. U. Tech, 2004 ]
[Hint: (JI "'"20, <12 "'"15 as the samples are large]
lO.A simple sample of heights of6400 Englishmen has a mean of67.85
inches and a s.d of 2.56 inches, while a simple sample of heights of 1600
Australians has a mean of 68.55 and a s.d of 2.52 inches. Do the
data indicate that Australians are on the average taller than the Englishmen?
2.33
Test at 1% level. Given J~(t )dt =.01.
o
[Hint : Since sample-size are so large we may take (JI "'" 2.56 ,
(J2 "'"252. Take HI(J.l1 < J.l2)]
l1.What is the test statistic which is used to test the equality of two
means of two normal populations 'when their common s.d. is unknown.
Answers
1. No ; /z/ = 25 > 1.96 2. No. Iz/ =3.33 3.0.065
11.
[Hint:Take Ho(P=~).Hl(P*t)]
9. In a sample of 400 burners there were 12 whose internal diameter's
were not within tolerance. Is this sufficient evi dence for concluding
that the manufacturing process is turning out more than 2% defective burners.
1·645
EM-2A-JJ
514 ENGlNEERING MATHEMATICS-IIA
[Give f = 0.05]
1~6
4> [Hint: Take Ho(P
6
=~)]
14. In a sample of 600 students of a certain college 400 are found to
use bi-cycle. In another college from a sample of 900 students 450 were
found to use bi-cycles. Test at 1% level whether the two colleges are
significantly different with respect to the habit of using bi-cycle. Given
area under standard normal curve enclosed between the ordinates z = -2.58
and z = -00 is 0·005.
15. Before an increase in excise duty on tea, 400 people out of a sample
of 500 persons were found to be tea drinkers. After an increase in duty,
400 people were tea drinkers in a sample of 600 people. State at 2.5%
level of significance whether there is a decrease in consumption of tea
o ,2
due to increment of excise duty. Given ~
,,2n
fe--
-1.96
2
dz=0.475.
16. In a year there are 956 births in town A, of which 52.5% were
males, while in town A and B combined, this proportion in a total of 1406
birth was 0.496. Is there any significant difference in the proportion of male
births in the two towns. Test at 5% level. See appropriate statistical table.
17. A machine produced 20 defective articles in a batch of 400. After
overhauling it produced 10 defectives in a batch of 300. Has the machine
improved? Test at 5% level of significance. Given area under standard
normal curve between the ordinates z = 1.645 and z = 00 is 0.05.
LARGE SAMPLE TEST OF SIGNIFICANCE 515
19.Two random samples of size 500 and 400 are drawn from two
normal populations having same s.d. 5. It is observed that the two sample
means are 11.5 and 10.9 respectively. Can the sample b~ regarded as drawn
from the same population. See appropriate table.
22.A company claims that its light bulbs are superior to those of a
competitor on the basis of a study which showed that a sample of 40 of
its bulbs had an average life time of 628 hours of continuous use with a
s.d of 27 hours, while a sample of 30 bulbs made by the competitor had
an average life time of 619 hours of continuous use with a s.d of 25 hours.
Check, at the 5% level of significance, whether this claim is justified.
23. The means of two large samples of size 1000 and 2000 are 67.5
and 68.0 respectively. Test at 1% level of significance whether the means
of the two population each with variance 6.25 are equal.
00 00
24. The mean yield of rice from a district A was 220kg per acre with
s.d 12kg recorded from a sample of 150 plots. In another district B,the
mean yield was 210kg with s.d 10kg per area from a sample of 100
plots.Assuming that the standard deviation of the yield in the entire state
was 11kg, test at 5% level whether there is any significance difference
between the mean yield of rice in the two districts.
[Give z025 == l.96, z.05 = 1.645 ]
25. LED bulbs manufactured by company A and company B gave the
following results:
company A company B
No. of bulbs used. 100 100
Average life in hour. 1248 1300
s.d in hour. 93 82
Find whether the average life of the two makes are same. Test at .05
level of significance
[Given area under standard normal curve enclosed between the ordinate
Z = 1.96 and Z =00 is .025]
26. A manufacturing company produces electric lights in each of its
two factories. It is suspected that the efficiency in the factories is not the
same. So a test is conducted and following datas are collected
Factor A Factor B
No of tube lights in sample 200 100
Average life 900 1100
s.d 220 240
27. The mean yield of two sets of plots and their variability are as
given below. Examine whether the difference in variability of yields is true.
Test at 1% level.
[Give z005 = 2.6 and ZOI = 2.3]
LARGE SAMPLE TEST OF SIGNIFICANCE 517
28. A potential buyer of elective bulbs bought 100 bulbs each of two
famous brands. Upon testing these he found that brand A had a mean life
of 1500 hours with a standard deviation of 50 hours whereas brand B
has a mean life of 1530hours with a standard deviation of 60 hours. Can
it be concluded at 1% level of significance that the two brands differ
significantly in quality? [Given zoos = 2.6, zOI = 2.3 ]
29. The mean height of 50 students who showed above average
participation in college atheletic was 68.2 inches with a standard deviation
of 2.5 inches; while 50 students who should no interest in such a
participation had a mean height of 67.5 inches with a standard deviation
of 2.8 inches. Test at 1% level the hypothesis that students who participate
in college atheletics are taller than other male students.
f rp=.OI][Hint:
00
33. In a sample of 600 families in a city, 400 families are found to use
broadband internet. In an another city, from a sample of 900 students 450
were found to use this type of internet. Test whether the two eities are
significantly different in using broadband internet.Test at 1% level
[Given z05 = 1.64; z.005 = 2.6]
Ansvvers
1. No.; z=-0.32 2. yes 3. No; Izl=3
4. No; Izl=2.37 5. No; Izl=2.67,
6. Izi = 8· 21; the population cannot have mean 5 em
7. Computed value of z = -2·597 . Supplier's claim is not valid
8. Computed value of z = 2·08 'The sexes are born not in equal
proportion
9. Value of z = 1·429, Manufacturing process is not turning out more
then 2% defective burners
10. value of z = -3·25 .Males and females in the population are not in 5
: 3 ratio
11. Value of z = 1·69 ; Rejected
12.Value of z = 2·27 , Company's claim is rejected
13.z = 1· 79 . The die is unbiased
14. yes; Izl=6.38 15. yes; z=4.94
16. z = 3.37; There is difference in A and B
17. not improved, z = 1.08
18. z=-3.9 Reject HO(I-Il =1-12)at 1% level
19. yes; z = 1.79 20. z = 1.86, No 21. yes, z = 1.16
23. Izl = 10.12 , not equal 24. Izl = 7.05, significantly different
25. Izl = 4.19 . Average life are not same 26. Izi = 0.96 . same
27. z = 1.3 , Not true, no difference.
28.lzl = 3.84 ;differ significantly 29. Izl = 1.32 ; not taller.
15.1. Introduction:
In the previous chapter test of significance / hypothesis was conducted
on the basis of large sample drawn from the population.In this chapter
this test will be done by drawing a sample of smaller size C < 30). Practically
it is convenient to draw a smaller random sample and sometimes a large
sample may not be available also.
In fact the process used in this direction is also a consequence of
some results discussed in the chapter of 'Sampling & Its Distribution' and
the well known 'Neyman pearson Theorem'.
·) - cr. X - Ilo ·f H . L ft id d
C1 X < Ilo - J;; Zu i.e., Z = cr/ J;; < -Zu 1 I 1S e S1 e .
··) -
C 11 X
cr.
> 1-.1.0 + J;; Zu i.e. Z
X -
= cr/
Ilo
J;; > Zu· 1if H I 1S
. Ri h id d
g t S1 e .
C···)
III
1- X - Ilo I > J;; . IIZ = Xcr/-Ilo
c zU/2 i.e, J;; > Zu/2 1if H I IS
. b ot I1 S1id e d ,
ote. Here if the population s.d a is not known we can not assume
CT the sample s.d
= S ,
In that case we go to the following method.
If the sample size is small « 30) and the sample is not drawn from
a normal population then the following theorem shows the method of
testing the hypothesis.
Theorem 2. ( To test Mean of any Population, not necessarily normal)
Let a random sample Xl ,x2,.· ,xn be taken from a population with mean
11 to test the null Hypothesis H 0 (11 = 110) against an alternative hypothesis
HI at a -level of significance. Then the Best Crilical Region determined
by the test statistic x (the sample mean) is given by
(iii) It I=
n-1 S
/~
> ta/2 if HI is both sided
Illustrative Example.
Example.I, A tyre manufacturing company claims that the average life
of their product is. 69 thousand miles. A car manufacturing company
planning to purchase tyre collect a sample of 10 lyre 's life from the
population of 'life' of tyres which are 65, 71, 64, 71, 70, 69, 64, 63, 67
and 68 thousands of miles. It is known that the population is normal and
the standard deviation of the population is .J7.056. Test at 5% level of
significance that the averege life of tyre of the company is lower than
they claim.
[Given Z.05 = 1.65, z.OI = 2.36, /'01;9 = 2.28]
Solution. We take the Null hypothesis H 0 (p = 69) and HI (p < 69) as the
alternative hypothesis.
Here Po = 69
Since population is normal we take the test statistic,
Z -
x - Po -
--
x-69
----;:::::==--;-- =--
x- 69
- a/ r: - .J7.056/ 0.84
/"n 1M
Here the sample mean,
x = 65 + 71 + 64 + 71 + 70 + 69 + 64 + 63 + 67 + 68 = 67.2
10
ENGINEERING MATHEMATICS -IIA
522
x x-45 = Y y2
50 5 25
49 4 16
52 7 49
44 -1 1
45 0 0
48 3 9
46 1 1
45 0 0
49 4 16
45 0 0
Total 23 117
:. y = 1~ :L>i = 1~ x 23 = 2 ·3
.: Y = x - 45 :. Y =x - 45 :. x = Y + 45 = 2·3 + 45 = 47·3
.Var (y) =~ LY; - (~LYi r = 1~ x 117 - (2.3)2 = 11. 7 - 5 . 29 = 6·41
47·3-50
Here the computed value of t = J9
= -3.20 with degrees offreedom
10-1=9. 2 ·531 9
t - Curve
-:2.26 2.26
I II III IV V
IQ before training: 110 120 123 132 125
-;
SMALL SAMPLE TEST OF SIGNIFICANCE 525
_ 10-2+2+4-4
The sample mean, x = =2
5
Calculation of sample s.d. :
10 -2 2 4 -4 10
100 4 4 16 16 140
X-Il
Here t = / r=; has n - I d.o.f
S -Jn=):
2-0
The computed value of t =
./ r: = O· 82
4·90/,,4
This is a two tailed test. Since t.01 = 4·6 we get the following figure:
t - curve
-4.6 4·6
The figure gives P(4. 6 < t < ex)) = ·01
By symmetry p( -00 < t < -4·6) = ·01 also
:. p(ltl > 4.6) = ·02.
:. the critical region is given by It I > 4·6 at 2% level of significance.
The computed value of t = 0·82 < 4·6 that is it does not lie in critical
Region. :. Ho is accepted at 2% level.
:. We conclude" there is no change in IQ"
15.3. Test for Difference of Mean (Test of equality of means)
Theoreml. (If the population is ormal) Let two independent random
samples {XI,X2,"'Xn1} and {x[,X2,,,,X:'2} be drawn from two Normal
of populations (111)0'1) and (112,0'2) respectively to test the hypothesis
:n
Ho(1l1 = 112) at a level of significance ;
is
XI =~(Xl
nl
+-X1 +".+xnJ
'
x 2 = ~(XI
n2 I
+ x' + ... + x' )
2 n2 .
ENGINEERING MATHEMATICS-11A
526
nl n2
statistic which is standard normal variate. Accordingly the C.R is
(i) z < -Za. if HI (Ill < 1l2) (left sided)
nl n2 (nl + n2 - 2) XI - X2
Hcre t= .~========
nl + n2 ~nlSf + n2Si
is the test statistic which is a students t variate with nl + n2 - 2 degre
of freedom where Sf ,Sf are variances of the two samples respectivel!
Accordingly the C.R is I - curve
Illustration.
Let two samples be drawn from two (Ilt>cr) and (1l2,cr) norm
populations. The details of the two samples are
mean s.d
size
7.52 0.024
6
7.49 0.032
5
SMALL SAMPLE TEST OF SIGNIFICANCE 527
We are to test the null hypothesis Ho(1l1 = 112)' o is not given, with
the help of the above sample-observations. We choose the test statistic
nln2 (nl + n2 - 2) XI - X2
t= '-r=======
+ n2 ~nIS? + n2S}
nl
which is a student-t variate with nl + n2 - 2 degrees of freedom.
Corresponding to the drawn sample the computed value of
6x5(6+5-2) 752-7.49
t =. = 160496
6+5 ~6 x (0.024)2 + 5 x (0.032)2 .
having d.o.f 6+5-2 = 9.
Let the alternative hypothesis be HI (Ill > 112), right sided.
From the statisticaltable we see p(t > 2.26) = 0.025 correspondingto d.o.f 9.
So at 0.025 level of significance the CR is t > 2.26 corresponding to
d.o.f 9. Here the computed value 1.60496 does not lie in this region. So
we accept Ho and conclude "the two means of the two populations are
equal" at 2.5% level of significance.
Exampe.1. There are two normal populations, population I and population
II. 8 and 6 are s.d of the two population respectively. Two independent
samples of size 10 and 12 are drawn from the two population respectively
with sample mean 20 and 27. Test at 1% level (i) whether the two
population mean are equal (ii) whether the population I mean is less than
the population II mean. [Take the help of statistical table}
Solution. If J.1.1 and J.1.2 be the population mean of the two population
then the null hypothesis is H 0 (J.1.1 = J.1.2) .
Since the population are normal we use the test statistic,
XI -x2
z=
a + __
_I a;
2
nl n2
Here XI =20, x2 =27, al =8, a2 =6, nl =10, n2 =12
20-27 -7
:. the computed value of z = -;====== or, z = -- = -2.28
82 62 3.07
-+-
20 12
528 ENGINEERING MATl-IEMATlCS-IIA
(i) Here consider the alternative hypothesis HI (J.il = J.i2) .Both sided.At
1 a
1% level a =- =.01 :. -= .005
100' 2
From the table we get za = z.005 = 2.58
2
00 -2.58
f rp = .005 f rp = .005 also
2.58
-----~
SMALL SAMPLE TEST OF SIGNIFICANCE 529
5 x 7 (5 + 7 - 2) 46 - 57
t = . = -1.70
5+7 v'5 x58+ 7 x 132.28
From the given data we get p(t < -2.2281) =.025
EM-2A-34
530 ENGINEERING MATHEMATICS-IIA
.. r~ n - 2. S d I • h •
at a Ieve I then t he test statistic, I= IS a tu ent s z-vanate WIt
1- r2
n - 2 degrees of freedom; r is the correlation -coefficients of the pairs in
the sample. Accordingly the CR is
(i) I < I_a if HI (p < 0) ,left sided
(ii) I > la if HI (p > 0) , right sided
hypothesis Ho(p = 0)
SMALL SAMPLE TEST OF SIGNIFICANCE 531
..
statistic 1
Z = -loge 1 + r h as approximate
-- . norma I diistnibuti . h
ution Wit
2 l-r
Z -Z
i.e. the test statistic z = 1__ 1_1 + _2_1_ is standard normal variate
V rtl- 3 n2 - 3
for which CR is obtained accordingly.
Example Two independent samples of sizes 16 and 18 are drawn from
r two normal bivariate population. It is observed that the correlation
coefficient between the variables in the two samples are 0.40 and 0.68
respectively. Test at 5% level of significance whether the correlation
coefficient of the variates in the two population are same.
[Given area under normal curve between the ordinates
z=-1.96 and z=1.96isO.95]
Solution. The sample sizes are nl = 16, n2 = 18 .The sample correlation
'1 = 0.40, r2 = 0.68 .Let PI' P2 be the correlation coefficients in the two
population respectively. The null hypothesis H 0 (p~ = P2) .
The alternative hypothesis HI (PI =#; P2) .Both sided.
I 1+r. 1 l+r.
ZI = -loge __ I and Z2 = -loge __ 2
2 l-rl 2 l-r2
1.96
Now s. =--S.
2n.
n. -1
2 13
=-x(1.8)
12
2
=3.51
.
2n2 2 9 2
=--S2
S2 =-x(2.1) =4.961
n2 -1 8
Computed value of test statistic,
F=s~ =~=0.708
s2 4.961
with degrees of freedom 13 -1, 9 -1 i.e. 12, 8.
2 n 2 6 2
Now SI =--SIl =-x(4.0) =19.20
nl -1 6-1
2 n2 2 5 2
S2 =--S2 =-x(2.1) =5.51
n2-1 5-1
19.20
= -T = --
s2
Computed value of F = 3.48
S2 5.51 •
with degrees of freedom (nl-l,n2 -1)=(5,4)
The CR is F > 6.26
Our computed value of F = 3.48 ~ 6.26
:. it does not lies in CR. Ho is accepted.
We conclude 0"1and 0"2 are equal.
v
.. computed value of t =
S
if
XI -X2
1
-+-
1
nl n2
Computation of 10 and Ie
Class Observed Probability Expected freqoercy
Frequency (under Ho) (Ie)
(10)
3
Students in H.S 200 3/8 - x 500 = 187.5
8
2
Students in 1st Year 150 2/8 -x500= 125
8
2
Students in 2nd Year 100 2/8 -x 500= 125
8
1
Students in 3rd Year 50 1/8 -x 500= 625
8
Total 500 1 500*
* The total of the expected frequencies should be equal to the total of
the observed frequencies.
2 ..., (10 _ j~)2 .
The test statistic is X = J'... Ie ' with 4 - 1= 3 degree of freedom.
The summation extends over all classes. Computation of X2
corresponding to the sample is shown below :
----
540 ENGINEERING MATHEMATlCS -IIA
J
SMALL SAMPLE TEST OF SIGNIFICANCE 541
Since our computed value does not lie in this region (as 7.92 '} 11.07)
we accept Ho at 5% level and conclude
No. of
12 8 20 2 14 10 15 6 9 4
accidents
Here k = No. of classes =; No. of weeks = 10 .
The accident conditions are same during these lO-week period means
there are equal number of accidents in each week.
~--------~~--~~-----~
542 ENGINEERING MATHEMATlCS-IiA
:. the CR is l > 16.919. Since the computed value lies in this region
we reject Ho at 5% level and conclude "The accident conditions are not
same during the 10-week period".
Ex. 4. Of 160 off springs of a certain cross between guinea pigs 102
were red, 24 were black and 34 were white. According to genetic model
the probabilities of red, black and white are respectively 9/16, 3/16 and
114. Test at 5% level of significance, if the data are consistent with the
model. For 2 degrees offreedom p( l > 5.99) = 0.05.
Here k = No. of type of guine-pig = 3. Take Ho (The probabilities of
red, black, white are 9116, 3/16 and 1/4 respectively); HI (the probabilities
are not so as prescribed in genetic model).
10 = observed frequency of the class in sample
Ie = expected frequency of the class in sample
J
S tALL SAMPLE TEST OF SIG IFICANCE 543
v
A.
2= .v:J,
L..
----'J._e ),-2 2
is X variate with 3-1 =2
2
d.o.f.
Computation of X
not lie in this region (as 3.7 j 5.99). Hence Ho is accepted at 5% level of
significance and we conclude " the data are consistent with the model".
Ex. 5. Survey of 320 families with 5 children each revealed the following
distribution:
No. of boys 5 4 3 2 1 0
No. of girls 0 1 2 3 4 5
No. of family 14 56 110 88 40 12
Is the result consistent with the hypothesis that male and female births
are equally probable? The 5% value of with 5 d.ofis 11.07. l
[WB.U.Tech, 2004, 2015
This is a sample of 320 families which are arranged as :
No. of boys 5 4 3 2 0
No.offamilies : 14 56 110 88 40 12
Here we think 'a family with 5 boys', 'a family with 4 boys' as the
classes. So here the number of classes, k =6.
We take Ho (the male and female birth are equiprobable) i.e. Ho (the
probability of male birth =..!.)
2
Then, among 5 children, the probability of r No. of boys
=5c·(..!.)r(1_..!.)5-r =_1 x5C
r 2 2 32 r •
544 ENGINEERING MATHEMATICS- IIA
The computed value of l = I (fa ~:e)2 = 7.16 with 5 d.o.f.. From the
supplied data we have the CR as l > 11.07 . Our computed value does
not lie in this region. So Ho is accepted at 5% level an~ we conclude "
1
the male and female birth are equi-probable".
Ex. 6. Four dice were thrown 112 times and the number of times 1, 3 or
5 was thrown were as under:
No. of dice throwing 0 2 3 4
a
1, 3, or 5
Frequency 10 25 40 30 7
Find the value of chi-square presuming that all dice were fair.
/
SMALL SAMPLE T~ST
,;.
OF SIGNlFlCANCE 545
_4 (1 )o( -1 )4-0_-- 1
- Co -
2 2 16
PI = Probability of 1 number of success in 4 trials
=4c{i)l(iY- = ~
1
. .
P2 = Probability of 2 succes in 4 trials' =4qGJ xGf2 = Ys
P3=4c{~J(~)= ~
4 (1)4(1)0 X6'
P4= C4"2 "2 =
r
546 ENGINEERING MATHEMATICS -IIA
J
SMALL SAMPLE TEST OF SIGNIFICANCE 547
Total frequency
In the above table, the expected frequency Ie of the category A2Bl
. 13x 27
IS~.
B A Total
- a fJ Rl
- Y 0 R2
Total Cl C2 N
J
548 ENGINEERING MATHEMATICS-IIA
Yate says
II 1
N { aO-2"a-2"0+~-f3r-2"f3-2"r-~
1 1 1 }2
=
R1R2 C1C2
=
k
N {( ao - f3r) - (a + f3+ r + o)} 2
R1R2 C1C2
N{(ao - f3r)-k~f
R1R2 C1C2
and for the case (ii)
2 N{(f3r-ao)-kNf
Corrected X = -~----~-
R1R2 C1Cz
That, considering both the case,
N{lao - f3r'-kNf
Corrected XZ --'----------"--
, R1R2'C1CZ
See the following Illustrative Examples.
--
/
SMALL SAMPLE TEST OF SIGNIFICANCE 549
-
SMALL SAMPLE TEST OF SIGNIFICANCE 55l
{
59 112x6-25x161-T
59}2
= = 7.44
37 x 22 x 28 x 31
with degrees of freedom (r -l)(c -1) = (2 -1)(2 -1) = 1
Given 5% value of X2 for 1 d.o.f is 3.84.
So the area under chi-curve (with 1 d.o.f) enclosed between the ordinates
X2 = 3.84 and X2 = 00 is .05
:. the CR is X2 > 3.84 at 5% level
The computed value of
X2 = 7.44 > 3.84
That is this falls in CR. So Ho is rejected at 5% level of significance.
We conclude' The vaccine has affect in controlling susceplibility of
tuberculosis
Exercise 15
1. Prices of shares of a company on the different days in a month
were found to be 66, 65, 69, 70, 69, 71, 70, 63, 64 and 68. Discuss
whether the average price of the shares in the month in 65
Can we conclude that the first population's mean is smaller than that of
second population. Use 1% level of significance. Given area under the
standard normal curve enclosed between the ordinates z = 0 and z = 2.33
is 0.49.
[Hint: 0,,02 are known; take Ho(1l1 = 112), H1(111 < 112) ]
11.A sample of size lO is drawn from each of two normal populations with '
the same unknown,variance and the followingresult are obtained:
Mean Variance ; the
Sample I 7 26
(the
Sample II 4 10
Test at 5% significance if the two populations have the same mean.
Given the following data. boys
Statistic d.o.f. Value at 5% level
t 18 2.101
z 1.96
554 ENGINEERiNG MATHEMATICS- lIA
12.A die is rolled 60 times and the following results are observed:
Face-point : 2 3 4 5 6 Total
No. of times: 6 10 8 13 11 12 60
occured
Are the data consistent with the hypothesis that the die is unbiased?
(Given "1.1, = 15.09 for 5 degrees of freedom.
1 -2.32_.e:..
[Given r;:::-
"l/2n
J
~
e 2 dJ-l = 0.01]
24. Two sets of ten students are selected at random from a college.
One set was given memory test without any training and the other set
was given the same test after two weeks of training. The scores obtained
are given in below:
Set A 10 8 7 9 8 10 9 6 7 8
Set B 12 8 8 10 8 11 9 8 9 9
Do you think: there is any significant effect of the training ?
(Given 1.05 ='2.10 at 8 d.o.f)
No. of heads
Frequency
°
14
1
45
2
80
3
112
4
61
5
8
Total
320
/
SMALL SAMPLE TEST OF SIGNIFICANCE 557
(i)Are the coins unbiased? Given X205 = 11.07 and X~1 = 15.09 for 5
degree of freedom.
(ii)Is the sample from a binomial population (given p(x2 > 11.07) = 0.05
for 5 d.o.f)
29.A sample analysis of certain examination results of 200 students in
a school was made. It was found that 46 students failed, 68 secured third
division, 62 secured second division and rest were placed in first division.
Are these results commensurate with the annual examination results which
are in the ratio of 2 : 3 : 3 : 2 for above various categories respectively?
(The tabulated value of chi-square for 3 d.o.f at 5% level of
significance is 7.81)
30. Test whether the following distribution fit the poisson distribution
with mean 0.6
x 0 1 2 3 4 5 or more
frequency 140 75 25 6 3 1
Given e-·6 = 05488, X205;3 = 7.81, X~5;4 = 9.49, X~5;5 = 11.07
31.A bird watcher sitting in a park has spotted a number of birds
belonging to 6 categories. The exact classification is given below.
Category: 1 2 3 4 5 6·
Frequency: 6 7 l3 17 6 5
Test at 5% level of significance whether or not the data is compatible
with the assumption that this particular park is visited by birds belonging
to these six categories in the proportion 1 : 1 : 2 : 3 : 1 : l. [Given
p(X2 > 11.07) = 0.05 for 5 d.o.f]
3:Z.The following table gives the number of aircraft accidents that
occured during the various days of the week. Test whether the accidents
are uniformly distributed over the week using 5% level of significance.
Day Sun Man Tue Wed Thu Fri Sat
No. of accidents l3 14 18 12 11 15 14
/
SMALL SAMPLE TEST OF SIGNIFICANCE 559
EM-2A-36
ENGINEERING MATHEMATlCS-IIA
562
Answer
1. t == 2.825 with 9 degrees of freedom "average price is not 65"
3. t == -4; Does not support
2. t == 3.l5
4. t == 1.555 with degrees of freedom 15; Ho is accepted
5. z == -2.28 . First population mean may not be smaller than second
population mean.
6. No. i.e. HI (PI ;I: P2) is rejected; t == 1.67
7. Yes, t == 4.99
8. No ; t == 2.04 . 9. t == -6.74, not same
21. t == 2 < 2.26 .The drug has no effect ill cLange of blood presure
o x
Figure. (i)
Figure. (ii)
t- curve
o
Figure. (iii)
F- curve F- cw·ve·
(5 percent F-points) (I percent F-points)
o F
Figure. (iv) Figure. (v)
STATISTICAL TABLES 565
x .00 .01 .02 .03 .04 .05 .06 .07 .08 .09
0.0 .5000 .4960 .4920 .4880 .4841 .4801 .4761 . .4721 .4681 .4641
0.1 .4602 .4562 .4522 .4483 .4443 .4404 .4364 .4325 .4286 .4247
0.2 . .4207 .4168 .4129 .4091 .4052 . .4013 .3974 .3936 .3897 .3859
0.3 .3821 .3783 .3745 .3707 .3669 .3632 .3594 .3557 .3520 .3483
0.4 .3446 .3409 .3372 .3336 .3300 .3264 .3228 3192 .3156 .3121
0.5 .3085 .3050 .3015 .2981 .2946 .2912 .2877 .2843 .2810 .2776
0.6 .2743 .2709 .2676 .2644 .2611 .2579 .2546 .2514 .2483 .2451
0.7 .2420 .2389 .2358 .2327 .2297 .2266 .2236 .2207 .2177 .2148
0.8 .2119 .2090 .2061 .2033 .2005 .1977 .1949 .1922 .1894 .1867
0.9 .1841 .1814 .1788 .1762 .1736 .1711 .1685 .1660 .1635 .1611
1.0 .1587 .1563 .1539 .1515 .1492 .1469 .1446 .1423 .1401 .1379
1.1 .1357 .1335 .1314 .1292 .1271 .1251 .1230 .1210 .1190 .1170
1.2 .1151 .1131 .1112 .1094 .1075 .1057 .1038 .1020 .1003 .0985
1.3 .0968 .0951 .0934 .0918 .0901 .0885 .0869 .0853 .0838 .0823
1.4 .0808 .0793 .0778 .0764 .0749 .0735 .0721 .0708 .0694 .0681
1.5 .0668 .0655 .0643 .0630 .0618 .0606 .0594 .0582 .0571 .0559
1.6 .0558 .0537 .0526 .0516 .0505 .0495 .0485 .0475 .0465 .0455
1.7 .0446 .0436 .0427 .0418 .0409 .0401 .0392 .0384 .0375 .0367
1.8 .0359 .0351 .0:\44 .0336 .0329 .0322 .0314 .0307 .0301 .0294
1.9 .0287 .0281 .0274 .0268 .0262 .0256 .0250 .0244 .0239 .02~~
2.0 .0228 .0222 .0217 .0212 .0207 .0202 .0197 .0192 .0188 .0183
2.1 .0179 .0174 .0170 .0166 .0162 .0158 .0154 .0150 .0146 .0143
2.2 .0139 .0136 .0132 .0129 .0125 .0122 .0119 .0116 .0113 .0110
2.3 .0107 .0104 .0102 .0099 .0096 .0094 .0091 .0089 .0087 .0084
2.4 .0082 .0080 .0078 .0075 .0073 .0071 .0069 .0068 .0066 .0064
2.5 .0062 .0060 .0059 .0057 .0055 .0054 .0052 .0051 .0049 .0048
2.6 .0047 .0045 .0044 .0043 .0041 .0040 .0039 .0038 .0037 .OOJ6
2.7 .0035 .0034 .0033 .0032 .0031 .0030 .0029 .0028 .0027 .0026
2.8 .0026 .0025 .0024 .oon. .0023 .0022 .0021 .0021 .0020 .0019
2.9 .0019 .0018 .0018 .0017 .0016 .0016 .0015 .0015 .0014 .0014
3.0. .0013 .0013 .0013 .0012 .0012 .0011 .0011 .0011 .0010 .0010
3.1 .0010 .0009 .0009 .0009 .0008 .0008 .0008 .0008 .0007 .0007
3.2 .0007 .0007 .0006 .0006 .0006 .0006 .0006 .0005 .0005 .0005
3.3 .0005 .0005 .0005 .0004 .0004 .0004 .0004 .0004 .0004 .0003
3.4 .0003 .0003 .0003 .0003 .0003 .0003 .0003 .0003 .0003 .0002
Table (ii).!l-Distribution
r~
n(/ .99 .98 .95 .90 .80 .70 .50 .30 .20 .10 .05 .02 .01 .001
.000 .001 .004 .016 '.064 .148
•
.020 .040 .455 1.074 1.642
.103 .211 .446 2.706
"I
3.940 4.865 6.179 14.684 16.919 19.679
7.267 9.342 11.781 21.666 27.877
13.442 15.987 18.307 21.161 23.209 29.588
3.609 4.575 5.578 6.989 8.148 10.341
12 3.571
'053 4.178 5.226 6.304 12.899 14.631 17.275
7.807 9.034 11.340 19.675 22.618 24.725
13 4.107 4.765 5.892 14.011 15.812 18.549 31.264
7.042 8.634 9.926 12.340 21.026 24.054
14 4.660 5.368 15.119 16.985 26.217 32.909
6.571 19.812
/I
7.790 9.467 10.821 22.362 25.472
15 5.229 5.985 13.339 16.222 18.151 21.064 27.688· 34,528
7.261 8.547 10.307 11.721 23.685 26.873
14.339 17.322 19.311 29.141 36.123
22.307 24.996 28.259 30.578 37.697
t'"
~
=
t'"
trl
[/)
rJ)
s;!
-l
Cii
-l
16 5.812 6.614 7.962 9.312 11.152 12.624 15.338 18.418 20.465 23.542 26.296 29.633 32.000 ~9.252 ?5
17 6.408 7.225 8.672 10.085 12.002 13.531 16.338 19.511 21.615 24.769 27.587 30.995 33.409 40.790 >
~
18 7.015 7.906 . 9.390 10.865 12.857 14.440 17.338 20.601 22.760 2.'i.989 28.869 32.346 34.805 42.312
8.567 10.117 11.651 13.716 15.352 18.338 21.689 23.900 27.204 30.144 :n.687 36.191 43.820 ~
19 7.633 co
20 8.260 9.237 10.851 12.443 14.578 16.266 19.337 22.775 25.038 28.412 31.410 35.020 37.566 -45.315 ~
l"l
rJ)
21 8.897 9.915 11.591 13.240 15.445 17.182 20.337 23.858 26.171 29.615 32.671 36.343 38.932 46.797
22 9.542 10.600 12.338 14.041 16.314 18.101 21.337 24.939 27.301 . 30.813 33.924 37.659 40.289 48.268
23 10.196 11.293 13.091 14.848 17.187 19.021 22.337 26.018 28.429 32.007 35.172 38.968 41.638 . 49.728
24 10.856 11.992 13.848 15.659 18.062 19.943 23.337 27.096 29.553 33.196 36.415 40.270 42.980 51.179
25 11.524 12.697 14·611 16.473 18.940 20.867 24.337 28.172 30.675 34.382 37.652 41.566 44.314 52.620
26 12.198 13.409 15.379 17.292 19.820 21.792 25.336 29.246 31.795 35.563 38.885 42.856 45.642 54.052
27 12.879 14.125 16.151 18.114 20.703 22.719 26.336 30.319 32.912 36.741 40.113 44.140 46.963 55.476
28 13.565 14.874 16.928 18.939 21.588 23.647 27.336 31.391 34.027 37.916 41.337 45.419 48.278 56.893
29 14.256 15.574 17.708 19.768 22.475 24.577 28.336 32.461 35.139 39.087 42.557 46.693 49.588 58.302
30 14.953 16.306 18.493 20.599 2\.364 2.'i.508 29.336 33.530 36.250 40.256 43.773 47.962 50.892 59.703
VI
I~
n (/ .45
Table (iii). rJ-Distribution
.40 .35 030 .25 .20 .15 .10 .05 .025 .01 .005 .0005
I /158./42
2
.325
.289
.510
.445
.727
.617
1.000 1.376 1.963 1078 6.314 /2.706
3 .137 .816 1.061 1386 31.82/ 63.657
.277 .424 .584 1.886 2.920 636.6/9
.765 .978 1.250 ' 4.303 6.965
4 .134 .271 .414 1.638 9.925 31.598
.569 .741 2353 3./82 4.541
5 .132 .267 .408 .941 1.190 1.533 5.841 12.924
.559 .727 2.132 2.776 ' 3.747
.920 1.156 1.476 4.604. 8.610
2.015 2.571
7
111 .130 .265 .404 .553 .718 .906 1.134
. 3.365 4.032 6.869
8 6/ .130
Y.l
.262
.402
.399
.549
.546
.7/1
.706
.896 1.1 19
1.440
1.415
1.943
/.895
2.447
2.365
3. 143 3.707 5.959
9 .129 .889 1.108 1.397 . 2.998 3.499 5.408
.26/ .398 .543 1.860 2.306
/0 ./29 .703 .883 1./00 2.896 30355
.260 .397 .542 1.383 /.833 5.041
.700 .879 2.262 2.821 3.250
/.093 1.372 1.812 4.781
2.228 2.764 1/69 4.587
" /129./28
/2
.260
.259
.396
.395
.540
.539
.697
.695
.876
.873
1.088 /.363 1.796 2.20/ 2.7/8
13 ./28 .259 /.083 1.356 1.782 3./06 4.437
.394
/4
15
.128
.128
.258
.258
.393
.393
.538
.537
.536
.694
.692
.691
.870
.868
.866
/.079
/.076
1.074
1.350
1.345
1.34/
1.771
1.761
1.753
2./79
2./60
2.145
2.13/
2.681
2.650
2.624
3.055
lOl2 .
2.977
4.3/8
4.22/
4.140
I ~.
~
I:n
-l
2.602 2.947 4.073 ?)
>
t""
~
Ctl
r-
t"l
f/j
--. _"1' _
r'-l
;?
16 .128 .258 .392 .535 .690 .865 1.071 1.337 1.746 2.120 2.583 2.921 4.014
-
-3
r'-l
-3
?}
17 .128 .257 .392 .534 .689 .863 1.069 1.333 1.740 2.110 2.567 2.898 3.965 >
.392 1.067 1.330 1.734 2.101 2.552 2.878 t""
18 .127 .257 .534 .688 .862 3.922
19 .127 .257 :391 .533 .688 .861 1.066 1.328 1.729 2.093 2.539 2.861 3.883 ~
20 .127 .257 .391 .533 .687 .860 1.064 1.325 1.725 2.086 2.528 2.845 3.850 =
t""
trI
r'-l
21 .127 .257 .391 .532 .686 .859 1.063 1.323 1.721 2.080 2.518 2.831 3.819
22 .127 .256 .390 .532 .686 .858 1.061 1.321 1.717 2.074 2.508 2.819 3.792
23 .127 .256 .390 .532 .685 .858 1.060 1.319 1.714 2.069 2.500 2.807 3.767
24 .127 .256 .390 .531 .685 .857 1.059 1.318 1.711 2.064 2.492 2.797 3.745
25 .127 .256 .390 .531 .684 .856 1.058 1.316 1.708 2.060 2.485 2.287 3.725
26 .127 .256 .390 .531 .684 .856 1.058 1.315 1.706 2.056 2.479 2.779 3.707
27 .127 .256 .389 .531 .684 .855 1.057 l.314 1.703 2.052 2.473 2.771 3.690
28 .127 .256 .389 .530 .683 .855 1.056 1.313 1.701 2.048 2.467 2.763 3.674
29 .127 .256 .389 .530 .683 .854 1.055 1.311 1.699 2.045 2.462 2.756 3.659
30 .127 .256 .389 .530 .683 .854 1.055 1.310 1.697 2.042 2.457 2.750 3.646
40 .126 .255 .388 .529 .681 .851 1.050 1.303 1.684 2.021 2.423 2.704 3.551
60 .126 .254 .387 .527 .679 .848 1.046 1.296 1.671 2.000 2.390 2.660 H60
120 .126 .254 .386 .526 .677 .845 1.041 1.289 1.658 1.980 2.358 2.617 3.373
00 .126 .253 .385 .524 .674 .842 1.036 1.282 1.645 1.960 2.326 2.576 . 3.291
IJI
e-
10
570
STATJSTJCAL TABLES
r'
2 3 4 5 6 8 /2 24 00
/99.5 2/5.7 224.6
2 /8.5/ 230.2 234.0 238.9 243.9
J9.00 /9./6 249.0
31 /0./3 9.55 9.28
/9.25 J9.30 /9.33 J9.37 '/9.4/ /9.45
254.3
J9.50
,r~
9./2 9.0/ 8.94
4 7.7/ 6.94' 6.59 8.84 8.74 8.64
6.39 6.26 6./6 8.53
5 6.6/ 5.79 5.4J 6.04 5.9/ 5.77
5./9 5.05 4.95 5.63
4..82 4.68 4.53 4.36
5./4 4.76 4.53
7 5..59 4.39 4.28
4.74 4.35 4./5 4.00 3.84
4./2 3.97 3.87 3.67
8 5.32 4.46 3.73 3.57 3.4/
4.07 3.84 3.69 3.23
9 5./2 3.58 3.44
4.26 3.86 3.63 3.48 3.28 3./2 2.93
10 4.96 3.37 3.23
4./0 3.7J 3.48 3.07 2.90 2.71
3.33 3.22 3.07 2.91 2.74 2.54
11
12 /4 ...
4.75
3.98
3.88
3.59
3.49
3.36
3.26
3.20 3.09 2.95 2.79 2.6/ 2.40
/3 4.67 3.80 3." 3.00 2.85
3.41 3./8 2.69 2.50 2.30
14 4.60 3.02 2.92 2.77
3.74 3.34 3.11 2.60 2.42
15 2.96 2.85 2.21
4.54 .l68 3.29 2.70 2.53 2.35
3.06 2.90 2.79 2.13
2.64 2.48 2.29 2.07
"/4.49
/7 4.45 3."
3.59
3.24
3.20
3.0/
2.96
2.85
2.8/
2.74 2.59 2.42 2.24 2.0/
/8 4.4/' 3.55 2.70 2.55 2.38
3.16 2.93 2.77 2.19 1.96
/9 4.38 3.52 2.66 2.5/ 2.34
3.n 2.90 2./5 1.92
20 4.35 3.49 2.74 . 2.63 2.48
3./0 2.87 2.3/ 2.1/ 1.88
2.7/ 2.60 2.45 2.28 2.08 1.84
3.47
22
23
211"2 4.30
4.28
3.44
3.42
3.07
3.05
2.84
2.82
2.68
2.66
2.57
2.55
2.42
2.40
2.25 2.05 1.8/
3.03 2.80 2.23 2.03 1.78
24 4.26 3.40 2.64 2.53 2.38
3.0/ 2.78 2.20 2.00 1.76
25 4.24 3.38 2.62 2.5/ 2.36
2.99 2.76 2.18 1.98 ·1.73
2.60 2.49 2.34 2.16 1.96 1.7/
2'/4.22
27 4.21
3.37
3.35
2.98
2.96
2.74
2.73
2.59
2.57
2.47
2.46
2.32 2./5 1.95 1.69
28 4.20 3.34 2.30 2.13 1.93
2.95 2.71 2.56 1.67
29 4.18 2.44 2.29
3.33 2.93 2.70 2./2 1.9/ 1.65
30 4.17 3.32 2.54 2.43 2.28
2.92 2.69 2.53 2./0 1.90 1.64
'2.42 2.27 2.09 1.89 1.62
40 /4.0'
60 4.00
3.23
3./5
2.84
. 2.76
2.6/
2.52
2.45
2.37
2.34 2./8 2.00 1.79 1.51
120 3.92 3.07 2.25 '2.10 1.92
2.68 2.45 1.70 1.39
3.84 2.29 2.17 2.02
00
2.99 2.60 1.83 1.61
2.37 2.2/ 2.10 1.25
1.94 1.75 1.52 1.00
'--
STATISTICAL TABLES 571
In
n 2 3 4 5 6 8 12 24 00
\ 4052 4999 5403 5625 5764 5859 5982 6\06 6234 6366
2 98.50 99.00 99.17 99.25 99.30 99.33 99.37 99.42 99.46 99.50
3 34.12 30.82 29.46 28.71 28.24 27.91 27.49 27.05 26.60 26.12
4 21.20 18.00 16.69 \5.98 15.52 15.2\ 14.80 14.37 13.93 \3.46
5 16.26 13.27 \2.06 \ \.39 10.97 10.67 10.29 9.89 9.47 9.02
6 \3.74 10.92 9.78 • 9.\5 8.75 8.47 8.\0 7.72 7.3\ 6.88
7 \2.25 9.55 8.45 7.85 7.46 7.19 6.84 6.47 6.07 5.65
8 \ 1.26 8.65 7.59 7.01 6.63 6.:n 6.03 5.67 5.28 4.86
9 10.56 8.02 6.99 6.42 6.06 5.08 5.47 5.11 4.73 4.31
\0 10.04 7.56 6.55 5.99 5.64 5.39 5.06 4.71 ·4.33 3.91
1\ 9.65 7.20 6.22 5.67 5.32 5.07 4.74 4.40 4.02 3.60
12 9.33 6.93 5.95 5.41 5.06 4.82 4.50 4.16 3.78 3.36
/3 9.07 6.70 5.74 5.20 4.86 4.62 4.30 3.96 3.59 3.16
\4 8.86 6.51 5.56 5.03 4.69 4.46 4.14 3.80 3.43 3.00
15 8.68 6.36 5.42 4.89 4.56 4.32 4.00 3.67 3.29 2.87
16 8.53 6.23 5.29 4.77 4.44 4.20 3.89 3.55 3.18 2.75
\7 8.40 6.11 5.18 4.67 4.34 4.10 3.79 3.45 3.08 2.65
\8 8.28 6.01 5.09 4.58 4.25 4.01 3.71 3.37 :tOO 2.57
\9 8J8 5.93 5.0\ 4.50 4.17 3.94 3.63 3.30 2.92 2.49
20 8.10 5.85 4.94 4.43 4.10 3.87 3.56 3.23 2.86 2.42
21 8.02 5.78 4.87 4.37 4.04 3.81 3.51 3.\7 2.80 2.36
22 7.94 5.72 4.82 4.31 3.99 3.76 3.45 3.12 2.75 2.31
23 7.88 5.66 4.76 4.26 3:94 3.71 3.41 3.07 2.70 2.26
24 7.82 5.61 4.72 4.22 3.90 3.67 3.36 3.03 2.66 2.2\
25 7.77 5.57 4.68 4:18 3.86 3.63 3.32 2.99 2.62 2.17
26 7.72 5.53 4.64 4.14 3.82 3.59 3.29 2.96 2.58 2.13
27 7.68 5.49 4.60 4.11 3.78 3.56 3.26 2.93 ~.55 2.10
28 7.64 5.45 4.57 4.m 3.75 353 3.23 2.90 2.52 2.06
29 7.60 5.42 4.54 4.04 3.73 350 3.20 2.87 2.49 2.03
30 7.56 5.39 4.51 4.02· 3.70 3.47 3.17 2.84 2.47 2.01
40 7.31 5.18 4.31 3.83 3.51 3.29 2.99 2.66 2.29 1.80
60 7.08 4.98 4.13 3.65 3.34 3.12 2.8.2 2.50 2.12 1.60
120 6.85 4.79 3.95 3.48 3.17 2.96 '2.66 2.34 1.95 1.38
00 6.64 4.60 3.78 3.32 3.02 2.80 2.51 2.18 1.79 1.00
INDEX 573