IME602Slides 01

IME602: PROBABILITY &
STATISTICS (Part 01)

Raghu Nandan Sengupta
Department of Management Sciences (DoMS)
Indian Institute of Technology Kanpur, INDIA
IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 1

(Part # 01)
Objective of the Course
 Nowadays the use of Probability and Statistics in all domains

of science and engineering is wide spread and ubiquitous.
 Hence it becomes imperative for the theoretician as well as
the practitioner to understand the basics and use probability
and statistical tools as well as methods in a well-grounded
manner.

(Part # 01)
Objective of the Course
(contd…)
 This course will undoubtedly benefit students in their masters and doctoral
programs when they work in a variety of areas like engineering, social
sciences, management science, etc., where probability, statistics and their
methodologies are used quite extensively.
 Furthermore the knowledge gained from this course will benefit learners
tackle and solve interesting problems both from theoretical as well as
practical viewpoints.

(Part # 01)
Key learning take aways
 Will help students master the rich repertoire of methodology and tools used in the
domain of Probability, Statistics, Distribution, Sampling Theory, Sampling
Distribution, Regression Analysis, etc.
 Will facilitate students with both the basic and advanced theoretical background in
Probability & Statistics.
 Will equip learners with the requisite skills in utilizing different techniques through
practical applications and solved examples.
• Will help participants learn and build on their expertise in the use of data, statistical
theory and statistical tools in myriad of applications like engineering, management
science, social science, basic sciences, etc.

(Part # 01)
Syllabus
Axioms of probability; Conditional probability; Discrete and continuous random
variables; Functions of random variables; Moments of random variables;
Generating functions; Limit theorems; Jointly distributed random variables;
Sufficiency and completeness; Descriptive and inferential Statistics; Sampling
theory and sampling distributions, Method for statistical inference; Theory of
point estimation and estimation of parameters; Theory of interval estimation;
Theory of hypotheses testing; Analysis of variance; Brief Introduction of
Multivariate Analysis; Linear and multiple linear regression; Introduction to
statistical Packages, e.g., MATLAB.

(Part # 01)
Text Books
1) Rohatgi, V. K. and Saleh, A. K. Md. E., An Introduction to
Probability and Statistics, John Wiley &Sons, 2001, ISBN
(10): 9814-12-603-9.
2) Larson, H. J., Introduction to Probability Theory and

Statistical Inference, John Wiley & Sons, 1982.
3) Ross, S. M., Introduction to Probability Models, Harcourt

Indian Private Ltd., 2000, ISBN (10): 0-12-598475-8.

(Part # 01)
Reference Books
1) Casella, G. and Berger, R. L., Statistical Inference, Duxbury Advanced
Series, 2002, ISBN (10): 81-315-0394-1.
2) Cox, D. R. and Hinkley, D. V., Theoretical Statistics, Chapman and Hall,

1974, ISBN (10): 0-412-12420-3.
3) Cràmer, H., Mathematical Methods of Statistics, Princeton University

Press, 1999, ISBN (13): 978-0-691-00547-8.
4) Feller, W., An Introduction to Probability Theory and its Applications

(Volume I), John Wiley & Sons, 2000, ISBN (10): 9971-51-315-3.

(Part # 01)
Reference Books (contd..)
5) Feller, W., An Introduction to Probability Theory and its Applications (Volume II),
John Wiley & Sons, 2000, ISBN (10): 9971-51-298-X.
6) Goon, A. M., Gupta, M. K. and Dasgupta, B., An Outline of Statistical Theory

(Volume I), The World Press Private Limited, Calcutta, 1998.
7) Goon, A. M., Gupta, M. K. and Dasgupta, B., An Outline of Statistical Theory

(Volume II), The World Press Private Limited, Calcutta, 2000, ISBN (10): 81-
87567-26-0.
8) Goon, A. M., Gupta, M. K. and Dasgupta, B., Fundamentals of Statistics (Volume

I), The World Press Private Limited, Calcutta, 2000, ISBN (10): 81-87567-25-2.

(Part # 01)
9) Goon, A. M., Gupta, M. K. and Dasgupta, B., Fundamentals of Statistics
(Volume II), The World Press Private Limited, Calcutta, 2001, ISBN (10): 81-
87567-18-X.
10)Hogg, R. V. and Craig, A. T., Introduction to Mathematical Statistics, Pearson

Education, 2004, ISBN (10): 81-7808-630-1.
11)Loeve, Michel, Probability Theory, Affiliated East West Press Pvt. Ltd., 1963.
12)Mood, A. M., Graybill, F. A. and Boes, D. E., Introduction to the Theory of

Statistics, Tata McGraw Hill Publication, 2001, ISBN (10): 0-07-044520-6.

(Part # 01)
13) Parzen, E., Modern Probability Theory and its Application, John Wiley & Sons,
1992, ISBN (10): 0471668257.
14) Lehmann, E. L. and Romano, J. P., Testing Statistical Hypotheses, Springer Verlag
Publishers, 2005, ISBN (10): 0387988645.
15) Draper, N. R. and Smith, H., Applied Regression Analysis, Wiley & Sons, 1981,
ISBN (10): 0471170828.
16) Walpole, R. E., Myers, R. H., Myers, S. L. and Ye, K., Probability and Statistics
for Engineers and Scientists, Pearson Education, 2007, ISBN (10): 81-317-1552-3.
17) Yule, G. U. and Kendall, M. G., An Introduction to the Theory of Statistics,

Universal Book Stall, 1999, ISBN (10): 81-85461-71-6.

(Part # 01)
Plan of classes (considering 14 weeks, with each
week 02 classes, of 1.25 hour each)
S.No. Coverage of Topics Week(s)
01 Axioms of probability; Conditional probability 01 to 02
02 Discrete and continuous random variables; Functions of random 03 to 04

variables; Moments of random variables
03 Generating functions; Limit theorems; Jointly distributed random 05 to 06

variables; Sufficiency and completeness
04 Descriptive and inferential Statistics; Sampling theory and 07 to 09

sampling distributions
05 Method for statistical inference; Theory of point estimation and 10 to 13
estimation of parameters; Theory of interval estimation; Theory of
hypotheses testing; Analysis of variance
06 Brief Introduction of Multivariate Analysis; Linear and multiple 14
linear regression
(Part # 01)
Evaluation Methodology
1) Quizzes: 20%
2) Assignments: 20%
3) Mid-term examination: 25%
4) Final examination: 35%

(Part # 01)
Software/Language
1) MATLAB <http://www.mathworks.com/> . One
can find sever based MATLAB at
https://www.iitk.ac.in/ccnew/
2) R <https://www.r-project.org/>
3) SPSS https://www.spss.co.in/
4) SAS <https://www.sas.com/en_in/home.html>

(Part # 01)
Practical cases
1) News paper vendor: wants to maximize
profit.
2) Production manager: wants to minimize
waiting time of jobs on machines and thus
reduce inventory and costs.
3) Marketing manager: wants to recommend
the best changes for a product (which is
being sold in the market) so that it will do
well.

(Part # 01)
Statistics
The word STATISTICS is derived
from the Italian word stato, which
means ″state″ and statista refers
to a person involved with the
affairs of the state.

(Part # 01)
Statistics
 Now a days, STATISTICS (in a plural
sense) is the study of qualitative and
quantitative data from our surrounding, be
it environment or any system so as to
draw meaningful conclusions about the
environment or system.
 It also means (in the singular sense) the
body of methods that are meant for
treatment of such data

(Part # 01)
Main steps in the study of
Statistics
 Method of collection of data
(primary or secondary)
 Scrutiny of data
 Presentation of data (non
frequency data, frequency
data)
(Part # 01)
Main steps in the study of
Statistics
 Analysis of data through
statistical models/methods
 Conclusions from results thus
obtained
 Modification of statistical
models/methods depending on
results obtained
(Part # 01)
Descriptive Statistics
Presentation of data
 Non frequency data
 Frequency data

(Part # 01)
Non frequency data
Time series or Historical data
Consider the case where the
representation of the values of one or
more variables like population of India,
price of petroleum etc., may be given for
different periods of time. For instance we
may be interested in knowing the
population change over time or the change
of production of petroleum over time.

(Part # 01)
BSE(30) Close
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
(Part # 01)
3-Jan-94
5-Jan-94
7-Jan-94
11-Jan-94
IME602:Probability & Statistics

13-Jan-94
17-Jan-94
19-Jan-94
24-Jan-94
27-Jan-94
31-Jan-94
2-Feb-94
4-Feb-94
8-Feb-94
10-Feb-94
14-Feb-94
16-Feb-94
Date
BSE(30) Close)
18-Feb-94
22-Feb-94
R.N.Sengupta,DoMS.,IIT Kanpur,INDIA
24-Feb-94
28-Feb-94
1-Mar-94
3-Mar-94
7-Mar-94
9-Mar-94
Non frequency data
15-Mar-94
17-Mar-94
21-Mar-94
23-Mar-94
25-Mar-94
Time series or Historical data
29-Mar-94
21
31-Mar-94
Non frequency data
Spatial series data
It may be that the values of one or more
variables are given for different individuals
in a group for the same period of time. But
instead of considering the group as such
we may be more interested in studying the
way the values of the variable(s) change
from individual to individual in that group.

(Part # 01)
Non frequency data
Spatial series data
Fertiliser Consumption for few Indian states for 1999-2000 (in tonnes)
Fertiliser consumption (in tonnes)
3500000
3000000
2500000
2000000
1500000
1000000
500000
Fertiliser Consumption States

(Part # 01)
Frequency data
We have the data on one or more variables for
different individuals for different periods of time
or for different points. But now we are more
interested in the characteristic(s) of the group
rather than the individuals in that group. In
studying the IQ level of students in a school we
may be interested in such group characteristics
as the percentage of students with IQ higher
than 130 or the percentage of students with
average IQ less than 90, etc.
(Part # 01)
Frequency data:
Tabular representation
India at a glance Year
(% of GDP) 1983 1993 2002 2003
Agriculture 36.6 31.0 22.7 22.2
Industry 25.8 26.3 26.6 26.6
Mfg 16.3 16.1 15.6 15.8
Services 37.6 42.8 50.7 51.2
Pvt Consump 71.8 37.4 65.0 64.9
GOI consump 10.6 11.4 12.5 12.8
Import 8.1 10.0 15.6 16.0
Domes save 17.6 22.5 24.2 22.2
Interests paid 0.4 1.3 0.7 18.3
Note: 2003 refers to 2003-2004; data are preliminary. Gross domestic
savings figures are taken directly from Indias central statistical
organization.

(Part # 01)
Frequency data
Diagrammatic representation
This is the most commonly used for representing time
series. The line diagram (also called histogram) is a
graph showing the relationship of the given variable with
time. There may be three types of line diagram, for the
scales used for both the axes of co-ordinates may be
arithmetic (or natural) scales, or one of them may be
arithmetic and the other logarithmic, or both may be
logarithmic. A line diagram where the vertical scale is
logarithmic but the horizontal scale is of the ordinary
arithmetic type is called a ratio chart or semi-logarithmic
chart. When both the vertical as well as the horizontal
axes are logarithmic then the chart is called the doubly-
logarithmic chart.
(Part # 01)
Frequency data
Line diagram representation
Fertiliser Consumption for few Indian states for 1999-2000 (in tonnes)
Fertiliser consumption (in tonnes)
3500000
3000000
2500000
2000000
1500000
1000000
500000
Fertiliser Consumption States

(Part # 01)
Frequency data: Bar diagram
(histogram) representation
World Population (projected mid 2004)
7000000000
6000000000
5000000000
Population
4000000000
3000000000
2000000000
1000000000
0
1950 1960 1970 1980 1990 2000
Year
World Population

(Part # 01)
World Population (projected mid 2004)
2000
1990
1980
Year
1970
1960
1950
0 1000000000 2000000000 3000000000 4000000000 5000000000 6000000000 7000000000
Population
World Population

(Part # 01)
Number of countries
12
10
8
Number
0
10 to 15 15 to 20 20 to 25 25 to 30 30 to 35 35 to 40 40 to 45 45 to 50
GDP in 1000 US$ (year 2002)

Number of countries

(Part # 01)
Height and Weight of individuals
200
180
160
140
Height/Weight
120
100
80
60
40
20
0
Ram Shyam Rahim Praveen Saikat Govind Alan
Individual
Height (in cms) Weight (in kgs)

(Part # 01)
Height and Weight of individuals
Alan
Govind
Saikat
Individual
Praveen
Rahim
Shyam
Ram
0 20 40 60 80 100 120 140 160 180 200
Height/Weight
Height (in cms) Weight (in kgs)

(Part # 01)
Frequency data
Pictorial diagram representation
In this type of presentation we can
represent the data more vividly and in
many a cases it is the popular method of
representing the data. Here a suitable
symbol is first chosen to represent a
definite number/quantities of units of the
variable. Against each data or observation
the symbol is represented proportionally
so that we can get the idea of the variable
quantities against that data point.
(Part # 01)
Frequency data
Pictorial diagram representation
Number of universities in the following states in USA
Note: Each of the  symbol represents 5 universities
Alabama 
Alaska 
Arizona 
Arkansas 
Colorado 
Connecticut 
Delaware 
Kansas 

(Part # 01)
Frequency data
Statistical map representation
If for example we are interested to show
diagrammatically regional seismicity in Alaska of
earthquakes of all magnitudes reported between
01/01/1960 to 11/09/2002. The colour code as
shown indicates the depth of the event. Thus
blue: 0 < h  33 km, green: 33 < h  75 km, red:
75 < h <= 125 km and yellow : h > 125 km. The
larger circles are earthquakes of M 7.0 and
higher from 1900-11/09/2002. The colour of
circle indicates the depth of the event, as above.
The star indicates the location of the 03/11/2002
(Part # 01)
Frequency data
Statistical map representation

(Part # 01)
Frequency data: Divided bar
diagram representation
Consider we have the time spent in hours
by student appearing for the CBSE
examination for the preparation of
Mathematics, Physics, Chemistry and
Biology. We collect the student's
preparation pattern for a five day period
and want to represent the data thus
obtained. In that case we would use the
divided bar diagram as illustrated below.
(Part # 01)
Frequency data: Divided bar
Time spent in preparation for subjects
8.0
7.0
6.0
5.0
Hours
4.0
3.0
2.0
1.0
0.0
Monday Tuesday Wednesday Thursday Friday
Day
Mathematics Physics Chemistry Biology

(Part # 01)
Frequency data: Stacked column
The method of depicting the data is almost
similar to the divided bar diagram
representation, but here we represent the
percentage wise figures for the variables for
each data point. Consider we are finding the
consumption in rupees for the four main
categories of food of a family in the months of
January to June. Remembering that the total
amount spent for each month can be different,
we depict the percentage wise consumption in
food for the four categories.
(Part # 01)
Frequency data: Stacked column
Percentage wise consumption of food
June
May
April
Month
March
February
January
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Rice Wheat Vegetables Cereals Percentage

(Part # 01)
Frequency data: Pie
diagram/chart representation
When the values of a variable are given
for a number of categories, as in spatial
series, we may be interested in a
comparison of the categories or series or
the contribution of each category to the
total. Here the proportions or percentages
of various categories, rather than the
absolute values for the categories, will be
the principal subject of study
(Part # 01)
Frequency data: Pie
diagram/chart representation
Median marks in JMET (2003)
Verbal Quantitative Analytical Data Interpresentation

(Part # 01)
Frequency data
Textual representation
In textual representation of data we depict the
information through text.
Consider for the year 2004-2005 we know the
number of post graduate students who have
registered in different engineering course at IIT
Kanpur. The figures are 83 in Aerospace, 88 in
Chemical, 139 in Civil, 222 in Electrical, 176 in
Mechanical, 115 in Computer Science. Given
this data we may be required to utilize this
information to answer some queries.
(Part # 01)
Frequency data
Stem leaf representation
The stem and leaf representation is a quick way of
looking at the data set. It contains the information of a
histogram but avoids the loss of information in a
histogram that results from aggregating the data into
intervals. The stem and leaf display is based on the
tallying principle but also uses the decimal base of our
number system. In the steam and leaf representation,
the stem is the number without its rightmost digit (the
leaf). The steam is written to the left of a vertical line
separating the steam from the leaf. Suppose we have
the numbers 105, 106, 107, 109, 100, 108. Then if we
use the steam and leaf representation we would depict
the numbers as 10 | 567908
(Part # 01)
Frequency data
Box plot representation
The box plot is also called the box
whisker plot. A box plot is a set of five
summary measures of distribution of the
data which are
 median
 lower quartile
 upper quartile
 smallest observation
 largest observation.

(Part # 01)
Frequency data
Whisker
LQ Median UQ

(Part # 01)
Frequency data
Here:
 UQ – LQ = Inter quartile range
(IQR)
 X = Smallest observation within
1.5(IQR) of LQ
 Y = Largest observation within
1.5(IQR) of UQ
(Part # 01)
Definitions
Quantitative variable: It can be described by a
number for which arithmetic operations such as
averaging make sense.
Qualitative (or categorical) variable: It simply
records a qualitative, e.g., good, bad, right, wrong, etc.
As already discusses statistics deals with

measurements, some being qualitative others being
quantitative. The measurements are the actual
numerical values of a variable. Qualitative variables
could be described by numbers, although such a
description might be arbitrary, e.g., good = 1, bad = 0,
right = 1, wrong = 0, etc.

(Part # 01)
Scales of measurement
Nominal scale: In this scale numbers are
used simply as labels for groups or classes. If
we are dealing with a data set which consists
of colours blue, red, green and yellow, then we
can designate blue = 3, red = 4, green = 5 and
yellow = 6. We can state that the numbers
stand for the category to which a data point
belongs. It must be remembered that nothing
is sacrosanct regarding the numbering against
each category. This scale is used for
qualitative data rather than quantitative data.
(Part # 01)
Ordinal Scale: In this scale of
measurement, data elements may be
ordered according to relative size or
quality. For example a customer or a
buyer can rank a particular
characteristics of a car as good, average,
bad and while doing so he/she can
assign some numeric value which may
be as follows, characteristic good = 10,
average = 5 and bad = 0.
(Part # 01)
Interval Scale: For the interval scale we specify intervals
in a way so as to note a particular characteristic, which
we are measuring and assign that item or data point
under a particular interval depending on the data point.
Consider we are measuring the age of school going
students between classes 5 to 12 in the city of Kanpur.
We may form intervals 10-12 years, 12-14 years,....., 18-
20 years. Now when we have one data point, i.e., the
age of a student we put that data under any one
particular interval, e.g. if the student's age is 11 years,
we immediately put that under the interval 10-12 years.

(Part # 01)
Ratio Scale: If two measurements are in
ratio scale, then we can take ratios of
measurements. The ratio scale
represents the reading for each recorded
data in a way which enables us to take a
ratio of the readings in order to depict it
either pictorially or in figures. Examples
of ratio scale are measurements of
weight, height, area, length etc.
(Part # 01)
Definitions
Population: Consists of the set of all measurements in
which the investigator is interested. Example can be all
the students in the city of Kanpur. The population is also
called the universe. A population is denoted by N
Sample: Is a subset of measurements selected from the

population. Sampling from the population is often done
randomly, such that every possible sample of n elements
will have an equal chance of being selected. A sample
selected in this way is called a simple random sample
or just a random sample. Example can be the students
in Kendriya Vidalaya inside IIT Kanpur campus. A
sample is denoted by n
(Part # 01)
Definitions
 Tally number: By tally number
we mean the tally we give
depending upon the number of
times that particular value of the
variable occurs in the total
universe or sample.

(Part # 01)
Definitions
 Frequency (absolute frequency): By
frequency (absolute frequency) we mean
the number of data points which fall within
a given class or for a given value in a
frequency distribution. It means that it
denotes the number of occurrence or
happening of a particular outcome. We
denote frequency (absolute frequency)
with fi.
(Part # 01)
Definitions
 Cumulative frequency: The cumulative
frequency corresponding to the upper
boundary of any class interval or value in a
frequency distribution is the total absolute
frequency of all values less (greater) than
that boundary for the class or value. We
denote cumulative frequency less (greater)
than type by F   f i ( F   f i )
n in n in

(Part # 01)
Example # 001
Consider we have the following data related to
the size in numbers of thirty families in the city of
Jaipur.
2, 6, 3, 4, 4, 5, 3, 6, 4, 4, 5, 3, 2, 3, 6, 5, 4, 4, 4,
3, 2, 4, 5, 6, 7, 4, 4, 5, 3, 3.
Now the question is how do we represent the

data using tally numbers, frequencies,
cumulative frequencies?
(Part # 01)
Example # 001 (contd..)
# of members Tally # fi F   fi F   fi
n n in
in
2 ||| 3 3 30
3 |||| || 7 10 27
4 |||| |||| 10 20 20
5 |||| 5 25 10
6 |||| 4 29 5
7 | 1 30 1

(Part # 01)
Pictorial representation of cumulative frequencies
35
30
25
Cumulative frequency
20
15
10
0
2 3 4 5 6 7
CF < then type CF < then type Number

(Part # 01)
Ogive and Median
A cumulative frequency diagram
is called the ogive. The abscissa
of the point of intersection
represents the median of the
data.

(Part # 01)
Guidelines for depicting
frequency tables
1) The classes should be mutually
exclusive and exhaustive
2) The number of classes should be neither
too small not too large
3) As far as practicable, the classes should
be of equal lengths

(Part # 01)
Example # 002
The following data relates to the height in cms of forty
individuals.
160.1, 167.2, 181.3, 154.7, 172.3, 161.3, 182.4, 158.2,
167.3, 159.4, 150.1, 157.3, 152.8, 155.8, 146.0, 162.0,
147.9, 149.9, 173.4, 166.4, 182.3, 151.2, 168.3, 170.1,
187.6, 163.4, 183.3, 171.9, 179.4, 166.8, 179.2, 168.3,
165.2, 166.7, 165.1, 166.3, 166.3, 173.4, 164.2, 164.9
1) We are required to prepare a frequency distribution table

showing the frequencies and the cumulative frequencies.
2) We are required to draw a histogram to exhibit the
frequency distribution graphically.
3) We are required to draw the two ogives.

(Part # 01)
Frequency table for the heights
Class Interval fi CF(< then) CF(>
then)
145.95-152.95 6 6 40
152.95-159.95 5 11 34
159.95-166.95 13 24 29
166.95-173.95 9 33 16
173.95-180.95 2 35 7
180.95-187.95 5 40 5

(Part # 01)
To draw the cumulative frequency less (greater)
than type, first specify the class intervals. Then
depict the class intervals along the horizontal
axis and the cumulative frequency less (greater)
than type along the vertical axis.
Against each class interval mark the point by the
corresponding cumulative frequency less
(greater) than type value. Join the points (the
values) depicting the cumulative frequencies
less (greater) than type with straight lines, which
will give us the respective ogives
(Part # 01)
Cumulative frequency chart
45
40
35
30
25
20
15
10
0
145.95-152.95 152.95-159.95 159.95-166.95 166.95-173.95 173.95-180.95 180.95-187.95
CF < then type CF < then type Class interval

(Part # 01)
Definitions
1) Class limit: The end points of any class
2) Class boundaries: The end points of any
class interval
3) Relative frequency: It is the ratio of f i
F
 Note: Using this concept of relative frequency

and cumulative relative frequency less
(greater) than type we can also draw the
cumulative relative frequencies curves

(Part # 01)
Definitions: Different
Measures
1) Measure of central tendency
 Mean (Arithmetic mean (AM), Geometric
mean (GM), Harmonic mean (HM))
 Median
 Mode
2) Measure of dispersion
 Variance or Standard deviation
 Skewness
 Kurtosis

(Part # 01)
Definition: Different Means
 Given N number of observations,
X1,….., Xn we define the following
AM=(X1+…..+Xn)/n
GM=(X1*…*Xn) 1/n
HM=n/(1/X1+…..+1/Xn)

(Part # 01)
Arithmetic Mean (m)
When estimating the long-term

expectation of a random variable,
the arithmetic mean is a natural
choice, e.g. finding the average
age of a group of persons,
average income of a group of
people etc.
(Part # 01)
Harmonic Mean
Consider a car travels a distance x with a
velocity of v1 and returns back the same
distance with a velocity of v2. What is the
average velocity?
2x
v
x x

v1 v2
Hence we see that in this case we use the

harmonic mean and not arithmetic mean

(Part # 01)
Harmonic Mean
Application areas are: (1) design stream
flow estimation for waste load allocation
(2) estimation of effective petrochemical
and geophysical properties of a
heterogeneous system of porous media.

(Part # 01)
Geometric Mean
Suppose you have an investment which
earns 10% the first year, 50% the second
year, and 30% the third year and we are
interested in finding an equivalent average
rate of return (geometric mean), say r:
Now we have
P(1+0.1)(1+0.5)(1+0.3) = P(1+r)3
Hence r = 28.97%
(Part # 01)
Median (e) and Mode (o)
 Median(e) : The median of a data set is
the value below which lies half of the data
points. To find the median we use F (e) =
0.5.
 Mode (o): The mode of a data set is the
value that occurs most frequently. Hence
f(o)  f(x);  x.

(Part # 01)
Example # 003

(Part # 01)
Example # 004

(Part # 01)
Example # 005
Consider we have the following data
points:
5, 7, 10, 7, 10, 11, 3, 5, 5
For these data points we have
m = 7; e = 10; o = 5; 2 = 6.89

(Part # 01)
Variance, Standard deviation,
Skewness, Kurtosis
 Variance: V[X] =   E X  E X 
2 2
 Standard deviation (SD) = 
3 3
 Skewness =  1    3
 3

  2 2
 4 
 Kurtosis =  2   2  3   4  3
 

(Part # 01)
Variance

(Part # 01)
Variance

(Part # 01)
Variance

(Part # 01)
Example # 006

(Part # 01)
Example # 007

(Part # 01)
Covariance & Correlation
Coefficient

(Part # 01)
Covariance

(Part # 01)
Covariance

(Part # 01)
Correlation Coefficient: Proof

(Part # 01)

(Part # 01)

(Part # 01)
Example # 008
143.995 147.122 148.456 148.884 150.028 150.245 151.375 151.465
Ht (cms) 140.421 1 2 1 1 5 6 3 6 151.789
62.4664 62.8498 64.2049 67.5864 68.2473 69.0382 69.8397 70.5646 70.6388
Wt (Kgs) 61.9169 9 1 3 3 1 2 2 6 3
152.174 152.599 156.278 160.920 161.346 161.542 166.054 168.003 168.222
Ht (cms) 1 8 3 3 1 3 1 9 9 171.456
71.8814 72.3498 73.8620 74.3739 79.7383 83.8073 84.9207 85.5075
Wt (Kgs) 3 3 73.7473 8 6 3 80.5299 8 5 9
•Given the height (in cms) and Weight (in Kgs) of 20

twenty students studying in Class XI at Sheiling House,
Kanpur. Find the mean, variance, correlation and
covariance values
•Remember this is a sample and NOT the population
hence we use the following terms
(Part # 01)
Example # 008 (contd…)

(Part # 01)
Descriptive statistics
Suppose the data are available in the form of a
frequency distribution. Assume there are k
classes and the mid-points of the corresponding
class intervals being x1, x2,…., xk. While the
corresponding frequencies are f1, f2,….., fk, such
that n = f1+f2+…..+fk
1 k 1 k 2
1
   xi f i   {  ( xi  x ) f i } 2
Then: n i 1 n i 1

(Part # 01)
Histogram for the heights
Frequency distribution
14
12
10
Frequency
0
145.95-152.95 152.95-159.95 159.95-166.95 166.95-173.95 173.95-180.95 180.95-187.95
frequency Class

(Part # 01)
Ogive curves for the heights
Cumulative frequency chart
45
40
35
30
25
20
15
10
0
145.95-152.95 152.95-159.95 159.95-166.95 166.95-173.95 173.95-180.95 180.95-187.95
CF < then type CF < then type Class interval

(Part # 01)
Descriptive statistics
Consider m groups of observations with
respective means 1, 2,….., m and standard
deviations 1, 2,….., m. Let the group sizes be
n1, n2,….., nm such that n = n1+ n2+…..+ nm.
Then: 1 m
 OVERALL    i f i
n i 1
1 m 2
m
2
1
 OVERALL  [ {  ni i   ni (  i   OVERALL ) }] 2
n i 1 i 1

(Part # 01)
Example # 009
In a batch of 10 children the IQ of the dull
boy is 36 below the average IQ of the
other children. Shown that the standard
deviation of IQ for all the children cannot
be less than 10.8. If the standard deviation
is actually 11.4, determine what is the
standard deviation when the dull boy is left
out.

(Part # 01)
Take k =2, such that n1 = 1 and n2 =9.
It is given that 1 - 2 = 36 and we also
know that 1 = 0. 2
1
2
 OVERALL  [0.9 2  116 .64]
Hence:
 From above we have OVERALL > 10.8
 If OVERALL = 11.4, we have 2 = 3.9.

(Part # 01)
Set Theory: Definition/Rules
 Set theory: Branch of
mathematical logic that studies
sets
 Set: A well defined collection of
objects, called numbers or
elements of the set

(Part # 01)

(Part # 01)

(Part # 01)

(Part # 01)

(Part # 01)

(Part # 01)

(Part # 01)

(Part # 01)

(Part # 01)

(Part # 01)
Example # 010

(Part # 01)
Venn Euler Diagram
U
A is Red
B is Yellow (Remember A is inside B)

(Part # 01)
Venn Euler Diagram
A is RED

(Part # 01)
Venn Euler Diagram
A B

(Part # 01)
Venn Euler Diagram

(Part # 01)
Venn Euler Diagram

(Part # 01)
Venn Euler Diagram

(Part # 01)
Venn Euler Diagram
U
A B
B-A=BAC

(Part # 01)
Venn Euler Diagram

(Part # 01)
Venn Euler Diagram

(Part # 01)

(Part # 01)

(Part # 01)

(Part # 01)

(Part # 01)
Probability
 Probability is the branch of mathematics
concerning numerical descriptions of how
likely an event is to occur, or how likely it is
that a proposition is true
 Probability of an event is defined as a
quantitative measure of uncertainty of the
occurrence of the event

(Part # 01)
Probability
Probability (P(A)): Of an event is defined as a quantitative
measure of uncertainty of the occurrence of the event
 Objective probability: Based on game of chance and

which can be mathematically proved or verified . If the
experiment is the same for two different persons, then the
value of objective probability would remain the same. It is
the limiting definition of relative frequency. Example: be
probability of getting the number 5 when we roll a fair die.
 Subjective probability: Based on personal judgment,
intuition and subjective criteria. Its value will change from
person to person. Example one person sees the chance of
India winning the test series with Australia high while the
other person sees it to be low.

(Part # 01)
Random event
Random experiment: Is an experiment whose
outcome cannot be predicted with certainty.
 Sample space () : The set of all possible
outcomes of a random experiment
 Sample point (i): The elements of the
sample space
 Event (A): Is a subset of the sample space
such that it is a collection of sample point(s).

(Part # 01)
Random event
For a random experiment, we denote
P(i) = pi P( A)   pi
 i A
Where:
 P(i) = pi = Probability of occurrence of the sample
point i
 P(A) = Probability of occurrence of the event
 P ()   p i  1
 i

(Part # 01)
Example #011
Suppose there are two dice each with faces 1, 2,....., 6
and they are rolled simultaneously. This rolling of the
two dice would constitute our random experiment
Then we have:
  = {(1,1), (1,2), (1,3), (1,4), (1,5), (1,6), (2,1),.…..,
(5,6), (6,1), (6,2), (6,3), (6,4), (6,5), (6,6)}.
 i = (1,1), (1,2),…., (6,5), (6,6)
 We define the event is such that the outcomes for
each die are equal in one simultaneous throw, then A
= {(1, 1), (2, 2),….., (6, 6)}
 P(i): p1 = p2 = ….. = p36 = 1/36
 P(A) = p1 + p8 + p15 + p22 + p29 + p36 = 6/36
(Part # 01)
Example # 012
Suppose a coin is tossed repeatedly till the first
head is obtained.
Then we have:
  = {(H), (T,H), (T,T,H),………}
 i = (H), (T,H), (T,T,H),…..
 We define the event such that at most 3 tosses
are needed to obtain the first head, then A =
{(H), (T,H), (T,T,H)}
 P(i): p1 = ½, p2 = (½)2, p3 = (½)3, p4 = (½)4,..…
 P(A) = p1 + p2 + p3 = 7/8
(Part # 01)
Classical definition of
probability
Under this definition we consider the
following:
 Sample space is finite.
 All the sample points are equally likely,
i.e., they have equal probability of
occurrence or equal relative frequency of
occurrence.

(Part # 01)
Example # 013
In a club there 10 members of whom5 are
Asians and the rest are Americans. A committee
of 3 members has to be formed and these
members are to be chosen randomly. Find the
probability that there will be at least 1 Asian and
at least 1 American in the committee
Total number of cases = 10C2 and the number of
cases favouring the formation of the committee
is 5C2*5C1 + 5C1*5C2
Hence P(A) = 100/120

(Part # 01)
Example # 014
There is box containing 10 coloured balls out of
which 4 are red, 4 are blue and 2 are white.
1) If you draw a ball at random what is the probability
that it is a white ball?
2) If you draw two balls consecutively and with
replacement then what is the probability that both
the balls are red?
3) If you draw two balls consecutively and without
replacement, then what is the probability that the first
is blue and the second is white?

(Part # 01)
Axiomatic definition of
probability
Under this definition we consider
the following:
 Sample space is infinite.
 Sample points are not equally
likely.

(Part # 01)
Example # 015
Suppose we continue with example 2
which we have just discussed and we
define the event B , that al least 5 tosses
are needed to produce the first head
  = {(H), (T,H), (T,T,H),………}
 i = (H), (T,H), (T,T,H),…..
 P(i): p1 = ½, p2 = (½)2, p3 = (½)3, p4 = (½)4,..…
 P(B) = p5+p6+p7+ ….. = 1 – (p1+p2+p3+p4)

(Part # 01)
Example # 016
Two fair coins are tossed simultaneously
and such simultaneously tossing is
repeated till you get two tails together.
What is the probability that you need at
most two such simultaneous tossing to
achieve our objective?

(Part # 01)
Probability as a Function
1
2
3
4
5 0 1/6
6
1
Domain Co-domain/Range

(Part # 01)
Probability, Chance, Relative
Frequency
 Consider you have the following readings for the speeds (km/hr) of 20 car
travelling past a crossing: 65, 66, 70, 64, 72, 70, 66, 77, 75, 65, 64, 69, 49,
53, 54, 55, 55, 70, 73, 66\
 Sum of relative frequency is 1 as it SHOULD be

(Part # 01)
Frequency (Unbiased)

(Part # 01)

(Part # 01)
Frequency (Biased)

(Part # 01)
Frequency (Biased)

(Part # 01)
Frequency (Biased)

(Part # 01)

(Part # 01)
Frequency (Biased)

(Part # 01)
Cumulative distribution function
(cdf) or the distribution function
We denote the distribution function by F(x)
F ( x)  P ( X  x)   f ( xi )
xi  x
x x
F ( x)  P( X  x)   f ( x)dx   dF ( x)
 

(Part # 01)
Properties of distribution
function
1) F(x) is non-decreasing in x, i.e.,
if x1 < x2, then F(x1)  F(x2)
2) Lt F(x) = 0 as x  - 
3) Lt F(x) = 1 as x  + 
4) F(x) is right continuous

(Part # 01)
Theorem in probability
For any event A, B  
 0  P(A)  1
 If A  B, then P(A)  P(B)
 P(A U B) = P(A) + P(B) – P(A  B)
 P(AC) = 1 – P(A)
 P() = 1
 P() = 0

(Part # 01)
Definitions
 Mutually exclusive: Consider n events
A1, A2,….., An. They are mutually exclusive
if no two of them can occur together, i.e.,
P(Ai  Aj) = 0.  i, j (i  j)  n
 Mutually exhaustive: Consider n events
A1, A2,….., An. They are mutually
exhaustive if at least one of them must
occur and P(A1UA2U…..UAn) = 1
(Part # 01)
Example # 017
Suppose a fair die with faces 1, 2,….., 6 is rolled. Then
 = {1, 2, 3, 4, 5, 6}. Let us define the events A 1 = {1,
2}, A2 = {3, 4, 5, 6} and A3 = {3, 5}
 The events A2 and A3 are neither mutually exclusive
nor exhaustive
 A1 and A3 are mutually exclusive but not exhaustive
 A1, A2 and A3 are not mutually exclusive but are
exhaustive
 A1 and A2 are mutually exclusive as well as
exhaustive
(Part # 01)
Unconditional Probability

(Part # 01)
Example # 018

(Part # 01)
Conditional probability
Let A and B be two events such that P(B) > 0.
Then the conditional probability of A given B is
P( A  B)
P( A | B) 
P( B)
Assume  = {1, 2, 3, 4, 5, 6}, A = {2}, B = {2, 4, 6}.
Then A  B = {2} and
16 1 16
P( A | B)   P( B | A)  1
36 3 16

(Part # 01)
Example # 019

(Part # 01)

(Part # 01)
Bayes Theorem
Let B1, B2,….., Bn be mutually exclusive
and exhaustive events such that P(Bi) > 0,
for every i =1, 2,….,
n n and A be any event
such that P( A)   P( A | Bi ) P(then
Bi ) we have
i 1
P( A | B j ) P( B j )
P( B j | A) 
n
 P( A | Bi ) P( Bi )
i 1

(Part # 01)
Bayes Theorem
B1 B2 B3 B4 B5
B6
B7 B8 B9
A
B10 B11

(Part # 01)
Baye’s Theorem: Proof

(Part # 01)
Example #020
In an examination each question has four
alternatives, answer of which only one is correct.
If a student knows the correct alternative then
he/she is definitely able to identify it. Otherwise
he/she picks one of the alternatives at random.
Given that a student has identified the correct
alternative what is the conditional probability that
he/she knew it, assuming 70% of the student
know the correct alternative to the question
under consideration.
(Part # 01)
Let us define the following events
 A1 = The student identifies the correct alternative
 B1 = The student knows the correct alternative
 B2 = The student does not know the correct
alternative
Then we know that P(B1) = 0.7, P(B2) = 0.3, P(A|
B1) = 1 and P(A|B2) = 0.25 from which we have
P(B1|A) = 0.9032

(Part # 01)
Example # 021
The marketing manager of a toy manufacturing company is
considering the marketing of a new toy. In the past 40% of
the toys introduced by the company have been successful
and 60% have been unsuccessful. Before the toy is
marketed, a market research is conducted and a report,
either favourable or unfavourable, is compiled. In the past,
80% of the successful toys received a favourable market
research report, and 30% of the unsuccessful received a
favourable market research report. The marketing manager
wants to know the probability that the toy will be successful
if it receives a favourable report

(Part # 01)
Independence of events

(Part # 01)
Example # 022
 Consider the example of rolling a die
 If A is the event the number appearing is odd and B be the
event the number appearing is a multiple of 3, then Pr(A)= 3/6
= and Pr(B) = 2/6
 Also A and B is the event the number appearing is odd AND a
multiple of 3 so that Pr(A ∩ B) = 1/6
 Pr(A│B) = Pr(A ∩ B)/ P(B)= (1/6)/(2/6) = 1/2 which implies
that the occurrence of event B has not affected the
probability of occurrence of the event A
 If A and B are independent events, then Pr(A│B) = Pr(A)
 Using Multiplication rule of probability, Pr(A ∩ B) = Pr(B) X
Pr(A│B) = Pr(B) X Pr(A)
(Part # 01)
Inequalities in Probability
 Inequality in probability theory are essential tools used
frequently in proofs.
 Probability Inequalities covers inequalities related with
events, distribution functions, characteristic functions,
moments and random variables (elements) and their sum.
 Examples
 Boole’s Inequality
 Bonferroni Inequality
 Markov’s Inequality
 Chebyshev’s Inequality
 Chernoff Bounds
 Cauchy-Schwarz Inequality

(Part # 01)
(contd…)
 Define the probability space
(Ω,ℱ,P).
Ω: Sample space
ℱ: Set of sub-sets of Ω
P: Function mapping of ℱ to
[0,1]
(Part # 01)
(contd…): Boole’s Inequality

(Part # 01)

(Part # 01)

(Part # 01)
Inequalities in Probability (contd…):
Bonferroni Inequality

(Part # 01)

(Part # 01)

(Part # 01)
Poincare’s Theorem

(Part # 01)
 Consider events A1 and A2, such that (A1A2) =
A1+A2-(A1A2) = A1 + {A2-(A1A2)}
 P(A1A2) = P(A1) + P{A2-(A1A2)} – P[{A1 {A2-
(A1A2)}]
 P(A1A2) = P(A1) + P{A2-(A1A2)} – P()
 P(A1A2) = P(A1) + P(A2) - P (A1A2), hence it
holds true for n=2
 Consider it is also true for n

(Part # 01)

(Part # 01)
Markov Inequality
 Let X be a non-negative r.v such
that E[X] exists
 Then for every positive a we
have
P(X  a)  E[X]/a

(Part # 01)
Distribution
Depending what are the outcomes of an
experiment a random variable (r.v) is used
to denote the outcome of the experiment
and we usually denote the r.v using X, Y
or Z and the corresponding probability
distribution is denoted by f(x), f(y) or f(z)
 Discrete: probability mass function (pmf)
 Continuous: probability density function
(pdf)
(Part # 01)
Discrete distribution
1) Uniform discrete distribution
2) Binomial distribution
3) Negative binomial distribution
4) Geometric distribution
5) Hypergeometric distribution
6) Poisson distribution
7) Log distribution

(Part # 01)
Bernoulli Trails
1) Each trial has two possible
outcomes, say a success and a
failure.
2) The trials are independent
3) The probability of success as well
as the probability of failure remains
the same from one trial to another

(Part # 01)
Uniform discrete distribution
[X ~ UD (a , b) ]
f(x) = 1/n x = a, a+k, a+2k,….., b
 a and b are the parameters where a, b  R
 E[X] = k
a (n  1)
2
 V[X] = 2 n 2  1
k ( )
 Example: Generating
12 the random numbers
1, 2, 3,…, 10. Hence X~UD(1,10) where
a=1, k=1, b=10. Hence n=10.

(Part # 01)
[X ~ UD (a , b) ]
0.12
0.1
0.08
0.06
f(x)
0.04
0.02
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

(Part # 01)
Binomial distribution
[X ~ B (p , n)]
f(x) = nCxpxqn-x x = 0, 1, 2,…..,n
 n and p are the parameters where p  [0, 1] and n  Z+

 E[X] = np
 V[X] = npq
 Example: Consider you are checking the quality of the
product coming out of the shop floor. A product can
either pass (with probability p = 0.8) or fail (with
probability q = 0.2) and for checking you take such 50
products (n = 50). Then if X is the random variable
denoting the number of success in these 50 inspections,
we have
X~50Cx(0.8)x(0.2)50-x
(Part # 01)
[X ~ B (p , n)]
0.16
0.14
0.12
0.1
0.08
f(x)
0.06
0.04
0.02
0
0 2 4 6 8 10 12 14 16 18 20 22 24 30 32 34 36 38 40 42 44 46 48 50
x

(Part # 01)
Example # 023
20% of the BSNL shares are owned
by Mr.Murali Lal. A random sample of
5 shares is chosen. What is the
probability that at most 2 of them will
be found to be owned by Mr.Murali
Lal?

(Part # 01)
Here p=0.2, q=0.8. Hence the required
probability is:
P(X  2) = P(X=0) + P(X=1) + P(X=2)
P(X  2) = 5C00.200.85 + 5C10.210.84 +
5
C2(0.2)2(0.8)3

(Part # 01)
Example # 024
Mr. Abhishek Bhatia finds that 70% of
the students have answered question
# 1 correctly in his chemistry
examination. A random sample of 10
answer scripts is chosen. What is the
probability that at least 3 of the
students have correctly answered the
question # 1?
(Part # 01)
Negative binomial distribution
[X ~ NB (p , r)]
f(x) = r+x-1Cr-1prqx x = r, r+1,.…..
 p and r are the parameters where p  [0, 1] and r  Z+
 E[X] = rq/p
 V[X] = rq/p2
 Example: Consider the example above where you are
still inspecting items from the production line. But now
you are interested in finding the probability distribution
of the number of failures preceding the 5th success of
getting the right product. Then, we have, considering
p=0.8, q=0.2
X ~ 5+x-1C5-1(0.8)5(0.2)x

(Part # 01)
[X ~ NB (p , r)]
0.35
0.3
0.25
0.2
f(x)
0.15
0.1
0.05
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
x

(Part # 01)
Example # 025
Suppose that the probability of a
manufacturing process producing a defective
item is 0.05. Suppose further that the quality
of any one item is independent of the quality
of any other item produced. If a Mr. Rao the
quality control officer selects items at random
from the production line and stops his
inspection when he gets 10 good items. Then
what is the probability of getting 2 defective
items before he stops inspection?
(Part # 01)
 Here p=0.95, q=0.05

 Hence the required
probability is P(X=2) =
10+2-1
C10-1(0.95) (0.05)
10 2

(Part # 01)
Example # 026
Mr. Agrawal is the immigration official at IGIA
and he knows that the probability of an Indian
origin person coming in any flight from abroad is
10%. He notes down the country of origin of
persons and stop when he knows he has noted
down exactly 3 people of Indian origin. Then
what is the probability of Mr.Agrawal noting
down 2 non Indian origin people before he stops
his checking?

(Part # 01)
Geometric distribution
[X ~ G (p)]
f(x) = pqx x = 0,1,2,…..
 p is the parameter where p  [0, 1]
 E[X] = q/p (r = 1 in the Negative Binomial distribution
case)
 V[X] = q/p2 (r = 1 in the Negative Binomial distribution
case)
 Example: Consider the example above. But now you
are interested in finding the probability distribution of
the number of failures preceding the 1 st success of
getting the right product. Then, we have considering
p=0.8, q=0.2
X~ (0.8)(0.2)x
(Part # 01)
[X ~ G (p)]
0.9
0.8
0.7
0.6
0.5
f(x)
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
x

(Part # 01)
Example # 027
A recent study indicates that Colgate toothpaste
has a market share of 45% (versus 55% of
Pepsodent). Miss. Dabawallah the marketing
research executive firm wants to conduct a new
taste test for which she wants users of Colgate.
Potential participants for the test are selected by
random screening of users of toothpaste to find
Colgate users. What is the probability that 5
users will have to be interviewed by Miss.
Dabawallah to find the 1st Colgate toothpaste
user?
(Part # 01)
Here p=0.45, q=0.55. Hence

the required probability is
P(X=4) = (0.45)1(0.55)4

(Part # 01)
Example # 028
Mr.Moti Singh is a lottery ticket sales
person and he knows that the
probability of a person coming to him
to check whether he/she has won the
lottery is 1%. What is the probability
that 10 persons arrive before the
winner arrives?

(Part # 01)
Hypergeometric distribution
[X ~ HG (N, n, p)]
f(x) = NpCxNqCn-x/NCn 0  x  Np and 0  (n – x)  Nq
 N, n and p are the parameters
 E[X] = np
 V[X] = npq{(N – n)/(N – 1)}
 Example: Consider the example above. But now you are interested in
finding the probability distribution of the number of failures(success) of
getting the wrong(right) product when we choose n number of products
for inspection out of the total population N. If the population is 100 and
we choose 10 out of those, then the probability distribution of getting the
right product, denoted by X is given by
X~85Cx15C10-x/100C10
 Remember
 p (0.85) and q (0.15) are the proportions of getting a good item and
bad item respectively.
 In this distribution we are considering the choosing is done without
replacement

(Part # 01)
[X ~ HG (N, n, p)]
0.4
0.35
0.3
0.25
f(x)
0.2
0.15
0.1
0.05
0
1 2 3 4 5 6 7 8 9 10
f(x)
x

(Part # 01)
Example # 029
Suppose that automobiles arrive at a Mr.
Ghoshs garage in lots of 10 and that for time
and resource considerations, he can inspect
only 5 out of each 10 for safety. The 5 cars
are randomly chosen from the 10 on the lot.
If 2 out of the 10 cars on the lot are below
standards for safety, what is the probability
that at most 1 out of the 5 cars to be
inspected by Mr.Ghosh will be found not
meeting the safety standards?
(Part # 01)
Here N=10, n=5, Np=2, Nq=8
Hence the required probability is
P(X1) = P(X=0) + P(X=1)
P(X1) = 2C08C5/10C5 + 2C18C4/10C5

(Part # 01)
Example # 030
Mrs. Patanaik is the teacher of class III consisting of
30 students. Daily during morning prayer session
she would like to check whether all her students
have brought their respective lunches. But due to
paucity of time each day she has to randomly select
10 students and finds that on an average out of this
10, 4 do not bring their lunches. If on any particular
day she selects 10 students then what is the
probability that exactly 5 students would not have
brought their respective lunches.

(Part # 01)
Poisson distribution
[X ~ P ()]
f(x) = e-x/x! x = 0,1,2,…..
  is the parameter where  > 0
 E[X] = 
 V[X] = 
 Example: Consider the arrival of the number of
customers at the bank teller counter. If we are
interested in finding the probability distribution of the
number of customers arriving at the counter in specific
intervals of time and we know that the average
number of customers arriving is 5, then, we have
X~ e-55x/x!

(Part # 01)
[X ~ P ()]
0.2
0.18
0.16
0.14
0.12
f(x)
0.1
0.08
0.06
0.04
0.02
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

(Part # 01)
Example # 031
Mr. Gurneek Singh who is the cashier at the
departmental store cash counter notices that
the average number of customer arriving at the
cash counter per 5 minutes is 10. Then what is
the probability that more than 5 customers
arrive at the cash counter with the interval of 5
minutes?

(Part # 01)
Here =10. Hence the required
probability is
P(X6) = 1 – {P(X=0) + P(X=1) +
P(X=2) + P(X=3) + P(X=4) + P(X=5)}
P(X6) = 1 – {exp(-10)100/0! + e-
10
101/1! + exp(-10)102/2! + exp(-
10)103/3! + exp(-10)104/4! + exp(-
10)105/5!}

(Part # 01)
Example # 032
Mr. Pankaj Yadav is the shop floor
manager and he notices that the average
number of jobs arriving at the lathe
machine per hour is 25. He wants to find
out what is the probability of exactly 10
jobs arrive in an hour at the lathe
machine?

(Part # 01)
Log distribution
[X ~ L (p)]
f(x) = -(loge p)-1x-1(1 – p)x x = 1,2,3,…..
 p is the parameter where p  (0, 1)
 E[X] = -(1-p)/(plogep)
 V[X] = -(1-p)[1 + (1 - p)/logep]/(p2logep)
 Example
1) Emission of gases from engines against fuel type
2) Used to represent the distribution of the number of
items of a product purchased by a buyer in a
specified period of time
(Part # 01)
Log distribution
[X ~ L (p)]
Log distribution
0.6
0.5
0.4
f(x)
0.3
0.2
0.1
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
x

(Part # 01)
Continuous distribution
1) Uniform distribution
2) Normal distribution
3) Exponential distribution
4) Chi-Square distribution
5) Gamma distribution
6) Beta distribution
7) Cauchy distribution

(Part # 01)
Continuous distribution
8) t-distribution
9) F-distribution
10)Log-normal distribution
11)Weibull distribution
12)Double exponential distribution
13)Pareto distribution
14)Logistic distribution

(Part # 01)
Uniform distribution
[X ~ U (a, b)]
f(x) = 1/(b – a) axb
 a and b are the parameters where a, b 
R and a < b
 E[X] = (a+b)/2
 V[X] = (b-a)2/12
 Example: Choosing any number between
1 and 10, both inclusive, from the real line

(Part # 01)
[X ~ U (a, b)]
0.12
0.1
0.08
f(x)
0.06
0.04
0.02
0
1 10
x

(Part # 01)
Normal distribution
[X ~ N ( , 2)]
( x   X )2
- < x <1  
2 X2
f ( x)  e
2  X
 X, 2X are the parameters where X  R and 2X

>0
 E[X] = X
 V[X] = 2X
 Example: Consider the average age of a student
between class VII and VIII selected at random
from all the schools in the city of Kanpur
(Part # 01)
Normal distribution
[X ~ N ( , 2)]
Normal distribution
0.25
0.2
0.15
f(x)
0.1
0.05
0
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
x

(Part # 01)
Standard normal distribution
Putting Z=(X-X)/X in the normal distribution we
have the standard normal distribution
z2
1 
f ( z)  e 2
2
Where: Z = 0 and Z = 1
Remember
• F(x) = P(X  x) = F(z) = F(Z  z)
• f(z) = (z) z z
• F ( z )  P( Z  z )   f ( z )dx   dF ( x)   ( z )
 
(Part # 01)
(PDF)
Standard Normal distribution
0.45
0.4
0.35
0.3
0.25
0.2
f(z)
0.15
0.1
0.05
0
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
-0.05
z

(Part # 01)
(CDF)
Normal F(x)
1.2
0.8
Normal F(x)
0.6
0.4
0.2
0
2.46
2.74
3.12
3.25
3.45
3.74
3.98
4.08
5.16
1.83
2.99
3.36
3.51
3.56
4.28
4.48
4.78
3.8
3.9
X

(Part # 01)
Normal distribution results
 (-z) =  (z), i.e., f(-z) = f(z)

 (-z) = 1- (z), i.e., P(Z  z) = 1 – P(Z
 z)
 P(a  X  b) = [(b - X)/X] - [(a -
X)/X]
 P(X  b) = [(b - X)/X]
 P(a  X) = 1 - [(a - X)/X]
(Part # 01)
Normal distribution results
Normal distribution
0.25
a b
0.2
0.15
f(x)
0.1
0.05
0
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
x

(Part # 01)
Finding Probabilities using Standard
Normal Distribution: P(0 < Z < 1.56)
Standard Normal Probabilities
z .00 .01 .02 .03 .04 .05 .06 .07 .08 .09
Standard Normal Distribution 0.0 0.0000 0.0040 0.0080 0.0120 0.0160 0.0199 0.0239 0.0279 0.0319 0.0359
0.1 0.0398 0.0438 0.0478 0.0517 0.0557 0.0596 0.0636 0.0675 0.0714 0.0753
0.4 0.2 0.0793 0.0832 0.0871 0.0910 0.0948 0.0987 0.1026 0.1064 0.1103 0.1141
0.3 0.1179 0.1217 0.1255 0.1293 0.1331 0.1368 0.1406 0.1443 0.1480 0.1517
0.4 0.1554 0.1591 0.1628 0.1664 0.1700 0.1736 0.1772 0.1808 0.1844 0.1879
0.3 0.5 0.1915 0.1950 0.1985 0.2019 0.2054 0.2088 0.2123 0.2157 0.2190 0.2224
0.6 0.2257 0.2291 0.2324 0.2357 0.2389 0.2422 0.2454 0.2486 0.2517 0.2549
0.7 0.2580 0.2611 0.2642 0.2673 0.2704 0.2734 0.2764 0.2794 0.2823 0.2852
f(z)
0.2 0.8 0.2881 0.2910 0.2939 0.2967 0.2995 0.3023 0.3051 0.3078 0.3106 0.3133
0.9 0.3159 0.3186 0.3212 0.3238 0.3264 0.3289 0.3315 0.3340 0.3365 0.3389
1.0 0.3413 0.3438 0.3461 0.3485 0.3508 0.3531 0.3554 0.3577 0.3599 0.3621
0.1
1.56 1.1 0.3643 0.3665 0.3686 0.3708 0.3729 0.3749 0.3770 0.3790 0.3810 0.3830
{
1.2 0.3849 0.3869 0.3888 0.3907 0.3925 0.3944 0.3962 0.3980 0.3997 0.4015
1.3 0.4032 0.4049 0.4066 0.4082 0.4099 0.4115 0.4131 0.4147 0.4162 0.4177
0.0
1.4 0.4192 0.4207 0.4222 0.4236 0.4251 0.4265 0.4279 0.4292 0.4306 0.4319
-5 -4 -3 -2 -1 0 1 2 3 4 5 1.5 0.4332 0.4345 0.4357 0.4370 0.4382 0.4394 0.4406 0.4418 0.4429 0.4441
Z 1.6 0.4452 0.4463 0.4474 0.4484 0.4495 0.4505 0.4515 0.4525 0.4535 0.4545
1.7 0.4554 0.4564 0.4573 0.4582 0.4591 0.4599 0.4608 0.4616 0.4625 0.4633
1.8 0.4641 0.4649 0.4656 0.4664 0.4671 0.4678 0.4686 0.4693 0.4699 0.4706
1.9 0.4713 0.4719 0.4726 0.4732 0.4738 0.4744 0.4750 0.4756 0.4761 0.4767
2.0 0.4772 0.4778 0.4783 0.4788 0.4793 0.4798 0.4803 0.4808 0.4812 0.4817
Look in row labeled 1.5 and 2.1

2.2
0.4821
0.4861
0.4826
0.4864
0.4830
0.4868
0.4834
0.4871
0.4838
0.4875
0.4842
0.4878
0.4846
0.4881
0.4850
0.4884
0.4854
0.4887
0.4857
0.4890
column labeled .06 to find

2.3 0.4893 0.4896 0.4898 0.4901 0.4904 0.4906 0.4909 0.4911 0.4913 0.4916
2.4 0.4918 0.4920 0.4922 0.4925 0.4927 0.4929 0.4931 0.4932 0.4934 0.4936
2.5 0.4938 0.4940 0.4941 0.4943 0.4945 0.4946 0.4948 0.4949 0.4951 0.4952
P(0 z 1.56) = .4406 2.6
2.7
0.4953
0.4965
0.4955
0.4966
0.4956
0.4967
0.4957
0.4968
0.4959
0.4969
0.4960
0.4970
0.4961
0.4971
0.4962
0.4972
0.4963
0.4973
0.4964
0.4974
2.8 0.4974 0.4975 0.4976 0.4977 0.4977 0.4978 0.4979 0.4979 0.4980 0.4981
2.9 0.4981 0.4982 0.4982 0.4983 0.4984 0.4984 0.4985 0.4985 0.4986 0.4986
3.0 0.4987 0.4987 0.4987 0.4988 0.4988 0.4989 0.4989 0.4989 0.4990 0.4990

(Part # 01)
z2
1 
Z ~ N(0,1) given by the equation f ( z)  e 2
2
The area within an interval (a,b) is given by
b z2
1  which is not integrable
F ( a  Z  b)   e 2 dz
a 2
algebraically. The Taylor’s expansion of the above assists in
speeding up the calculation, which is
1 1  (1) k z 2k 1
F (Z  z)   
2 2 k  0 (2k  1)2 k k!

(Part # 01)

(Part # 01)
Example # 033
In an examination 20% of the students failed
(i.e., obtained a score which is less than or
equal to 40 marks out of 100) and 10% of the
students obtained a grade A (score of 70
marks of above out of 100). Assuming normal
distribution of marks find the mean and the
standard deviation of the distribution

(Part # 01)
Steps
1) P(X  40) = 0.2 = P[(X-X)/X  (40- X)/X] = P(Z 
z1) = (z1) = -0.84
2) P(X  70) = 0.1 = P[(X-X)/X  (70- X)/X] = P(Z 
z2) = 1 - P(Z  z2) = 1 - (z2). Hence (z2) = 0.9
Hence we have from the above two equations:
 z1 = (40 - X)/X = - 0.84
 z2 = (70 - X)/X = + 0.90
 X = 54.12; X = 17.64
(Part # 01)
Example # 034
In Prof. Ram Pals mathematics examination
20% of the students failed (i.e., obtained a
score which is less than or equal to 40 marks
out of 100) and 10% of the students obtained
a grade A (score of 70 marks of above out of
100). Assuming normal distribution of marks
find the mean and the standard deviation of
the distribution of marks in mathematics?

(Part # 01)
Example # 034
1) P(X  40) = 0.2 = P[(X-X)/X  (40- X)/X] = P(Z  z1)
= (z1) = -0.84
2) P(X  70) = 0.1 = P[(X-X)/X  (70- X)/X] = P(Z  z2)

= 1 - P(Z  z2) = 1 - (z2). Hence (z2) = 0.9
Hence we have from the above two equations:
 z1 = (40 - X)/X = - 0.84
 z2 = (70 - X)/X = + 1.28
 X = 51.9; X = 14.2

(Part # 01)
z .00 .01 .02 .03 .04 .05 .06 .07 .08 .09
0.0 0.0000 0.0040 0.0080 0.0120 0.0160 0.0199 0.0239 0.0279 0.0319 0.0359
0.1 0.0398 0.0438 0.0478 0.0517 0.0557 0.0596 0.0636 0.0675 0.0714 0.0753
0.2 0.0793 0.0832 0.0871 0.0910 0.0948 0.0987 0.1026 0.1064 0.1103 0.1141
0.3 0.1179 0.1217 0.1255 0.1293 0.1331 0.1368 0.1406 0.1443 0.1480 0.1517
0.4 0.1554 0.1591 0.1628 0.1664 0.1700 0.1736 0.1772 0.1808 0.1844 0.1879
0.5 0.1915 0.1950 0.1985 0.2019 0.2054 0.2088 0.2123 0.2157 0.2190 0.2224
0.6 0.2257 0.2291 0.2324 0.2357 0.2389 0.2422 0.2454 0.2486 0.2517 0.2549
0.7 0.2580 0.2611 0.2642 0.2673 0.2704 0.2734 0.2764 0.2794 0.2823 0.2852
0.8 0.2881 0.2910 0.2939 0.2967 0.2995 0.3023 0.3051 0.3078 0.3106 0.3133
0.9 0.3159 0.3186 0.3212 0.3238 0.3264 0.3289 0.3315 0.3340 0.3365 0.3389
1.0 0.3413 0.3438 0.3461 0.3485 0.3508 0.3531 0.3554 0.3577 0.3599 0.3621
1.1 0.3643 0.3665 0.3686 0.3708 0.3729 0.3749 0.3770 0.3790 0.3810 0.3830
1.2 0.3849 0.3869 0.3888 0.3907 0.3925 0.3944 0.3962 0.3980 0.3997 0.4015
1.3 0.4032 0.4049 0.4066 0.4082 0.4099 0.4115 0.4131 0.4147 0.4162 0.4177
1.4 0.4192 0.4207 0.4222 0.4236 0.4251 0.4265 0.4279 0.4292 0.4306 0.4319
1.5 0.4332 0.4345 0.4357 0.4370 0.4382 0.4394 0.4406 0.4418 0.4429 0.4441
1.6 0.4452 0.4463 0.4474 0.4484 0.4495 0.4505 0.4515 0.4525 0.4535 0.4545
1.7 0.4554 0.4564 0.4573 0.4582 0.4591 0.4599 0.4608 0.4616 0.4625 0.4633
1.8 0.4641 0.4649 0.4656 0.4664 0.4671 0.4678 0.4686 0.4693 0.4699 0.4706
1.9 0.4713 0.4719 0.4726 0.4732 0.4738 0.4744 0.4750 0.4756 0.4761 0.4767
2.0 0.4772 0.4778 0.4783 0.4788 0.4793 0.4798 0.4803 0.4808 0.4812 0.4817
2.1 0.4821 0.4826 0.4830 0.4834 0.4838 0.4842 0.4846 0.4850 0.4854 0.4857
2.2 0.4861 0.4864 0.4868 0.4871 0.4875 0.4878 0.4881 0.4884 0.4887 0.4890
2.3 0.4893 0.4896 0.4898 0.4901 0.4904 0.4906 0.4909 0.4911 0.4913 0.4916
2.4 0.4918 0.4920 0.4922 0.4925 0.4927 0.4929 0.4931 0.4932 0.4934 0.4936
2.5 0.4938 0.4940 0.4941 0.4943 0.4945 0.4946 0.4948 0.4949 0.4951 0.4952
2.6 0.4953 0.4955 0.4956 0.4957 0.4959 0.4960 0.4961 0.4962 0.4963 0.4964
2.7 0.4965 0.4966 0.4967 0.4968 0.4969 0.4970 0.4971 0.4972 0.4973 0.4974
2.8 0.4974 0.4975 0.4976 0.4977 0.4977 0.4978 0.4979 0.4979 0.4980 0.4981
2.9 0.4981 0.4982 0.4982 0.4983 0.4984 0.4984 0.4985 0.4985 0.4986 0.4986
3.0 0.4987 0.4987 0.4987 0.4988 0.4988 0.4989 0.4989 0.4989 0.4990 0.4990

(Part # 01)
Example # 035
Dr. Satish Kashyap finds that his patients
height are normally distributed with mean
165 cms and standard deviation 20 cms.
What is the probability of a patient height
being between 160 to 170 cms?

(Part # 01)
Exponential distribution
[ X ~ E (a, )]
( x  a)
1 
f ( x)  e 
 a<x<
 a and  are the parameters where a  R
and  > 0
 E[X] = a + 
 V[X] = 2
 `
Example: The life distribution of the
number of hours a electric bulb survives.
(Part # 01)
[ X ~ E (a, )]
0.3
0.25
0.2
0.15
f(x)
0.1
0.05
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
x

(Part # 01)
(CDF)
Exponential F(x)
0.018
0.016
0.014
0.012
0.01
F(X)
0.008
0.006
0.004
0.002
0
1.83
2.46
2.74
2.99
3.12
3.25
3.36
3.45
3.51
3.56
3.74
4.08
4.78
5.16
3.98
4.28
4.48
3.8
3.9
X

(Part # 01)
Example # 036
Mr. K.Bharadwaj an electrician knows that
the time a bulb operates before it gets
fused is exponentially distributed with a
mean life of 100 hours. What is the
probability that any bulb will work
continuously for at least 200 hours?

(Part # 01)
Here  = 100. hence the required
probability is
P(X200) = 1 – P(X200)
1 200
x
P( X  200)  1   exp( )dx
x 0 100 100

(Part # 01)
Example # 037
Mr. Lyndoh the owner of an electronics
shop knows that the average life of a
tape recorder he sells is 1.5 years.
What is the probability that any such
tape recorder would function for at
most 3 years?

(Part # 01)
Example # 038 (Electronic
Components)
In a quality control survey the following data was obtained from the log book of QC, which
gives the number of hours of working before the electronic component fails and the
number of such components
Number of hours Number failed Number of hours Number failed
of working after this hour of working after this hour
2.0 41 6.0 27
5.0 30 4.0 34
3.5 35 1.0 45
1.5 43 2.5 39
4.5 32 7.0 25
0.5 48 6.5 26
10.0 18 5.5 29
3.0 37 7.5 24
8.5 21 8.0 22
9.5 19 9.0 20

(Part # 01)
Components) (contd…)
Frequency of number of failures
60
50
Number of failures
40
30
20
10
0
0.0 2.0 4.0 6.0 8.0 10.0
Frequency Hours of w orking

(Part # 01)
Probability
0.0900
0.0800
0.0700
0.0600
Probability
0.0500
0.0400
0.0300
0.0200
0.0100
0.0000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Reading number
P[X=x]=f(x)

(Part # 01)
Cumulative Probability
1.0000
0.8000
Cumulative Probability
0.6000
0.4000
0.2000
0.0000
0.0 2.0 4.0 6.0 8.0 10.0
Cumulative Probability Hours of w orking

(Part # 01)
 Probability that the electronic component
will survive after 6 hours?
 Average life of the electronic component?
 Standard deviation of life for the electronic
component?
 Probability that the electronic component
will fail before 4 hours?

(Part # 01)
Example # 039 (NO2)
URL: http://lib.stat.cmu.edu/datasets/
URL: http://lib.stat.cmu.edu/datasets/NO2.dat
The data are a subsample of 500 observations from a data set
that originate in a study where air pollution at a road is related to
traffic volume and meteorological variables, collected by the
Norwegian Public Roads Administration. The response variable
(column 1) consist of hourly values of the logarithm of the
concentration of NO2 (particles), measured at Alnabru in Oslo,
Norway, between October 2001 and August 2003. The predictor
variables (columns 2 to 8) are the logarithm of the number of
cars per hour, temperature 2 meter above ground (degree C),
wind speed (meters/second), the temperature difference
between 25 and 2 meters above ground (degree C), wind
direction (degrees between 0 and 360), hour of day and day
number from October 1, 2001.

(Part # 01)
Example # 039 (NO2) (contd…)
Temp(25)- Wind
Tem(2 directi
Ln(No2) ln(Cars) Temp(2 m) Wind Speed ) on Hour of day Day #
Average 3.698368 6.973342 0.8474 3.056 0.1494 143.3704 12.382 310.474
SD(S) 0.750597 1.087166 6.524636 1.784172 1.065237 86.51021 6.802693 200.9778
SD(P) 0.749846 1.086079 6.518108 1.782387 1.064171 86.42366 6.795887 200.7767
Median 3.84802 7.42536 1.1 2.8 0 97 12.5 212
Maximum 6.39509 8.34854 21.1 9.9 4.3 359 24 608

(Part # 01)
Ln(No2) Concentration
7
6
Ln(No2) Concentration
5
4
3
2
1
0
1
26
76
51
101
126
151
176
201
226
251
276
301
326
351
376
401
426
451
476
Data #
Ln(No2)

(Part # 01)
ln(Cars)
9
8
7
6
ln(Cars)
5
4
3
2
1
0
101
176
301
326
401
426
451
26
51
76
126
151
201
226
251
276
351
376
476
1
Data #
ln(Cars)

(Part # 01)
Temp(2 m)
25
20
15
10
Temp (2 m)
5
0
1
27
53
79
105
131
157
183
209
235
287
313
339
391
417
443
469
261
365
495
-5
-10
-15
-20
-25
Data #
Temp(2 m)

(Part # 01)
Wind Speed
12
10
8
Wind Speed
0
1
26
51
76
126
151
176
201
226
251
301
326
351
376
401
426
451
476
101
276
Data #
Wind Speed

(Part # 01)
Temp(25)-Tem(2)
4
Temp(25)-Tem(2)
0
1
53
27
79
131
287
313
365
105
157
183
209
235
261
339
391
417
443
469
495
-2
-4
-6
Data #
Temp(25)-Tem(2)

(Part # 01)
Wind direction
400
350
300
Wind direction
250
200
150
100
50
0
1
79
27
53
131
157
105
183
209
235
261
287
313
339
365
391
417
443
469
495
Data #
Wind direction

(Part # 01)
ln(Cars) vs Ln(No2)
9
8.5
8
7.5
7
Ln(Cars)
6.5
6
5.5
5
4.5
4
1 2 3 4 5 6 7
Ln(No2)

(Part # 01)
Temp(2m) vs Ln(No2)
22
20
18
16
Temp(2m)
14
12
10
8
6
4
1.5 2 2.5 3 3.5 4 4.5 5
Ln(No2)

(Part # 01)
Frequency distribution of Ln(No2)
180
160
140
120
Frequency
100
80
60
40
20
0
2 to 2.5
5.5 to 6
1 to 1.5
1.5 to 2
2.5 to 3
3 to 3.5
3.5 to 4
4 to 4.5
4.5 to 5
5 to 5.5
6 to 6.5
Data Range
Frequency

(Part # 01)
Frequency Distribution of Ln(No2)
100
90
80
70
Frequency
60
50
40
30
20
10
0
1 to 1.25
3 to 3.25
2 to 2.25
4 to 4.25
5 to 5.25
6 to 6.25
5.5 t 5.75
1.5 to 1.75
2.5 to 2.75
3.5 to 3.75
4.5 to 4.75
Data Range
Frequency

(Part # 01)
Cumulative Relative Frequency of Ln(No2)
1.00000
0.90000
Cumulative Relative Frequency of
0.80000
0.70000
0.60000
Ln(No2)
0.50000
0.40000
0.30000
0.20000
0.10000
0.00000
0
10
15
20
25
Ln(No2) Range

(Part # 01)
Example # 040 (Arrival times at CNC
work station)
In a factory shop floor for a certain CNC machine (machine
marked # 1) the number of jobs arriving per unit time are given
below
# of Arrivals Frequency
0 2
1 4
2 4
3 1
4 2
5 1
6 4
7 6
8 4
9 1
(Part # 01)
work station) M/C # 01 (contd…)
Frequency distribution of arrivals
7
5
Frequency
0
0 1 2 3 4 5 6 7 8 9
# of Arrivals

(Part # 01)
Relative Frequency of Number of Arrivals
0.25
0.2
Relative Frequency
0.15
0.1
0.05
0
0
9
# of Arrivals

(Part # 01)
Cumulative Relative Frequency of Number of Arrivals
1
0.9
Cumulative Relative Frequency
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
10
# of Arrivals

(Part # 01)
1) The probability of number of arrivals of
jobs being equal to or more than 7 is
about 0.18.
2) The average number of arrival of jobs is
5.
jobs being less than or equal to 4 is
about 0.45.
(Part # 01)
In the same factory, same shop floor, for the same type of CNC
machine (machine marked # 2) the number of jobs arriving per unit
time are given below. Here we consider machine # 2 is being used
for some other type of different machining operation
# of Arrivals Frequency
4 8
5 11
6 19
7 15
8 14
9 22
10 18
11 15
12 8

(Part # 01)
Frequency distribution of arrivals
25
20
15
Frequency
10
0
4 5 6 7 8 9 10 11 12
# of Arrivals

(Part # 01)
Relative Frequency of Number of Arrivals
0.18
0.16
0.14
Relative Frequency
0.12
0.1
0.08
0.06
0.04
0.02
0
4
10
12
11
# of Arrivals

(Part # 01)
Cumulative Relative Frequency of # of Arrivals
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
4
10
11
12
# of Arrivals

(Part # 01)
jobs being equal to or more than 7 is
about 0.60.
2) The average number of arrival of jobs is
8.
jobs being less than or equal to 4 is
about 0.05.
(Part # 01)
Example # 040 (Arrival times at CNC work
station) M/C # 01 and M/C # 02 (contd…)
1) Can you comment anything about the
bottleneck machine considering that the
processing time in both machine # 1 and
machine # 2 is same?
2) Can these sort of concepts be extended for
more than two machines?
3) What more things should we consider, apart
from processing time in each machine, in order
to find the optimal production rate per unit time
(say per shift)?

(Part # 01)
Example # 041 (SAT)
This data set [ ] includes eight variables:
1) STATE: Name of state
2) COST: Current expenditure per pupil (measured in thousands of
dollars per averagedaily attendance in public elementary and
secondary schools)
3) RATIO: Average pupil/teacher ratio in public elementary and
secondary schools during Fall 1994
4) SALARY: Estimated average annual salary of teachers in public
elementary and secondary schools during 1994-95 (in thousands of
dollars)
5) PERCENT: percentage of all eligible students taking the SAT in
1994-95
6) VERBAL: Average verbal SAT score in 1994-95
7) MATH: Average math SAT score in 1994-95
8) TOTAL: Average total score on the SAT in 1994-95

(Part # 01)
Example # 041 (SAT)
(contd…)
Histogram for Cost
12
10
COST State

(Part # 01)
Example # 041 (SAT)
(contd…)
Histograms for Cost and Ratio
30
25
Cost and Ratio
20
15
10
5
0
Nevada
Illinois
Texas
Maine
Georgia
Oregon
Virginia
Alabama
Michigan
New Mexico
Missouri
North Dakota
South Carolina
Arkansas
Kansas
Connecticut
Wisconsin
COST RATIO State

(Part # 01)
Cost, Ratio and Salary
(Part # 01)
10
20
30
40
50
60
0
Alabama
Arizona
California
COST
IME602:Probability & Statistics

Connecticut
RATIO
Florida
Hawaii
Illinois
SALARY
Iowa
Kentucky
Maine
Massachusetts
Minnesota
Missouri
State
Nebraska
New
Histogram of Cost, Ratio and Salary
R.N.Sengupta,DoMS.,IIT Kanpur,INDIA
New Mexico
(contd…)
North Carolina
Ohio
Oregon
Rhode Islan
South Dakota
Example # 041 (SAT)
Texas
Vermont
Washington
Wisconsin
262
Example # 041 (SAT)
(contd…)
1 n
Average value is given by E X    X i
n i 1
1 n
V  X    X i  E X 
Variance is given by 2
n i 1
Covariance is given by
CovX , Y   E X  E X Y  E Y    X ,Y V X  V Y 

(Part # 01)
Example # 041 (SAT)
(contd…)
COST RATIO SALARY PERCENT VERBAL MATH TOTAL
Mean 5.90526 16.858 34.82892 35.24 457.14 508.78 965.92
Median 5.7675 16.6 33.2875 28 448 497.5 945.5
Maximum 9.774 24.3 50.045 81 516 592 1107
Minimum 3.656 13.8 25.994 4 401 443 844
Standard Deviation(1) 1.362807 2.266355 5.941265 26.76242 35.17595 40.20473 74.82056
Standard Deviation(2) 1.34911 2.243577 5.881552 26.49344 34.82241 39.80065 74.06857

(Part # 01)
Example # 041 (SAT)
(contd…)
COST 1.820097 -1.12303 6.901753 21.18202 -19.2638 -18.7619 -38.0258
RATIO -1.12303 5.033636 -0.01512 -12.6639 4.98188 8.52076 13.50264
SALARY 6.901753 -0.01512 34.59266 96.10822 -97.6868 -93.9432 -191.63
PERCENT 21.18202 -12.6639 96.10822 701.9024 -824.094 -916.727 -1740.82
VERBAL -19.2638 4.98188 -97.6868 -824.094 1212.6 1344.731 2557.331
MATHS -18.7619 8.52076 -93.9432 -916.727 1344.731 1584.092 2928.822
TOTAL -38.0258 13.50264 -191.63 -1740.82 2557.331 2928.822 5486.154

(Part # 01)
Example # 041 (SAT)
(contd…)
COST 1 -0.37103 0.869802 0.592627 -0.41005 -0.34941 -0.38054
RATIO -0.37103 1 -0.00115 -0.21305 0.063767 0.095422 0.081254
SALARY 0.869802 -0.00115 1 0.61678 -0.47696 -0.40131 -0.43988
PERCENT 0.592627 -0.21305 0.61678 1 -0.89326 -0.86938 -0.88712
VERBAL -0.41005 0.063767 -0.47696 -0.89326 1 0.970256 0.991503
MATHS -0.34941 0.095422 -0.40131 -0.86938 0.970256 1 0.993502
TOTAL -0.38054 0.081254 -0.43988 -0.88712 0.991503 0.993502 1

(Part # 01)
Example # 042 (S&P)
URL: http://lib.stat.cmu.edu/datasets
URL: http://lib.stat.cmu.edu/datasets/spdc2693
The data set consists of the Standard and Poor's 500
Index closing values from 1926 to 1993. See also
djdc0093. Submitted by eduardo ley,
(edley@eco.uc3m.es) [13/Mar/96] (333 kbytes)
Standard and Poor's 500 Index closing values from 1926
to 1993.
The first column contains the date (yymmdd), second
column contains the value. These data are used in:
E.Ley (1996): "On the Peculiar Distribution of the U.S.
Stock Indices;" forthcoming in The American Statistician
(Part # 01)
Example # 042 (S&P)
(contd…)
Value of S&P
25
20
Value (USD)
15
10
0
Date
VALUE Date

(Part # 01)
Example # 042 (S&P)
(contd…)
For our problem we take n=300
 E(X)= 17.09927
 SD(Sample)=2.649796
 SD(Population)=2.645376

(Part # 01)
Example # 042 (S&P)
(contd…)
Frequency Distribution
80
70
60
Frequency
50
40
30
20
10
0
Frequency Range

(Part # 01)
Example # 042 (S&P)
(contd…)
Relative Frequency Distribution
0.3
0.25
Relative Frequency
0.2
0.15
0.1
0.05
0
11 to 12
13 to 14
14 to 15
15 to 16
16 to 17
17 to 18
18 to 19
19 to 20
20 to 21
12 to 13
Relative Frequency Range

(Part # 01)
Example # 042 (S&P)
(contd…)
1
0.8
0.6
0.4
0.2
0
0 2 4 6 8 10
Range Number

(Part # 01)
Log-normal distribution
[X ~ LN ( , 2)]
(loge x   X ) 2
f ( x) 
1 1
e

2 X2 0<x<
2  X x
 X, 2X are the parameters where X  R

and 2X > 0
 E[X] = exp(X+2X/2)
 V[X] = exp(2X+2X){exp(2X)-1}
 Example: Stock prices return distribution
(Part # 01)
[X ~ LN ( , 2)]
0.012
0.01
0.008
0.006
f(x)
0.004
0.002
0
0.5 3 5.5 8 10.5 13 15.5 18 20.5 23 25.5 28 30.5 33 35.5 38
x

(Part # 01)
Relationship between Poisson
and Exponential distribution
If a process has the intervals
between successive events as
independent and identical and it
is exponentially distributed then
the number of events in a
specified time interval will be a
(Part # 01)
Few approximations
Normal approximation to Binomial
distribution
Let X ~ B(p, n) where n is large and p
is small. Then the distribution can be
approximated by the Normal
distribution X ~ N(np, npq)

(Part # 01)
De Moivre Laplace Limit
Theorems
b  np a  np
1)
P( a  X  b)  [ ]  [ ]
npq npq
a  np
2) P (a  X )  1  [ ]
npq
b  np
3) P ( X  b)  [ ]
npq

(Part # 01)
Tchebychevs Inequality
 Let X be a r.v such that E[X] = X
and V[X] = 2X
 Then for every positive t we have
P(|X - X|  tX)  1/t2

(Part # 01)
Bernoullis Theorem
 Let Xn be the number of success
in n number of Bernoulli trials,
each with success probability p
 Then for arbitrary positive  we
have
Lt P[|Xn/n – p|)  ] = 1 as n  
(Part # 01)
Central Limit Theorem

(Part # 01)
Convergence Theorems
 In probability theory, there exist different
notions of convergence of random
variables.
 Convergence of sequences of random
variables to some limit random variable
is an important concept in probability
theory/applications to statistics/stochastic
processes.

(Part # 01)
(contd…)
 Stochastic convergence (in general mathematics)
formalizes the idea that a sequence of essentially random
or unpredictable events can sometimes be expected to
settle down into a behavior that is essentially unchanging
when items far enough into the sequence are studied.
 Convergence is related to some fixed/constant value, or
increasing preference towards a certain outcome or
increasing aversion against straying far away from a
certain outcome.
 Convergence is related to some unchanging
probability distribution as values in the sequence
continue to change.

(Part # 01)
(contd…)
 Convergence in distribution
 Convergence in probability
 Almost sure convergence
 Sure convergence/Point wise
convergence
 Convergence in mean
 Other concepts of convergence
(Part # 01)
(contd…)
 We assume that {X1,X2,….} is a
sequence of random variables, and
X is a random variable, and all of
them are defined on the same
probability space (Ω,ℱ,P).
Ω: Sample space
ℱ: Set of sub-sets of Ω
P: Function mapping of ℱ to [0,1]
(Part # 01)
Convergence Theorems (contd…):
Convergence in Distribution

(Part # 01)
Convergence in Distribution

(Part # 01)
Convergence in Probability

(Part # 01)

(Part # 01)
 Convergence in probability is a
condition on the joint cdf's, as
opposed to convergence in
distribution, which is a
condition on the individual cdf's

(Part # 01)
(contd…)
 Big O notation is a mathematical notation
that describes the limiting behavior of a
function when the argument tends towards
a particular value or infinity. It also means
grows no faster than
 Small o is a mathematical notion that
describes the upper bound of a function. It
also means grows strictly slower than

(Part # 01)
Convergence Almost Surely

(Part # 01)
Convergence Surely

(Part # 01)
Convergence in Mean

(Part # 01)
(contd…)

(Part # 01)

IME602Slides 01

Uploaded by

Copyright:

Available Formats

You might also like

IME602Slides 01

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

IME602Slides 01

Uploaded by

Copyright:

Available Formats

IME602: PROBABILITY &

STATISTICS (Part 01)

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 1

 Nowadays the use of Probability and Statistics in all domains

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 2

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 3

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 4

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 5

2) Larson, H. J., Introduction to Probability Theory and

3) Ross, S. M., Introduction to Probability Models, Harcourt

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 6

2) Cox, D. R. and Hinkley, D. V., Theoretical Statistics, Chapman and Hall,

3) Cràmer, H., Mathematical Methods of Statistics, Princeton University

4) Feller, W., An Introduction to Probability Theory and its Applications

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 7

6) Goon, A. M., Gupta, M. K. and Dasgupta, B., An Outline of Statistical Theory

7) Goon, A. M., Gupta, M. K. and Dasgupta, B., An Outline of Statistical Theory

8) Goon, A. M., Gupta, M. K. and Dasgupta, B., Fundamentals of Statistics (Volume

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 8

10)Hogg, R. V. and Craig, A. T., Introduction to Mathematical Statistics, Pearson

12)Mood, A. M., Graybill, F. A. and Boes, D. E., Introduction to the Theory of

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 9

17) Yule, G. U. and Kendall, M. G., An Introduction to the Theory of Statistics,

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 10

S.No. Coverage of Topics Week(s)

01 Axioms of probability; Conditional probability 01 to 02

02 Discrete and continuous random variables; Functions of random 03 to 04

03 Generating functions; Limit theorems; Jointly distributed random 05 to 06

04 Descriptive and inferential Statistics; Sampling theory and 07 to 09

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 12

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 13

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 14

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 15

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 16

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 19

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 20

IME602:Probability & Statistics

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 22

Fertiliser Consumption States

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 23

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 25

Fertiliser Consumption States

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 27

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 28

0 1000000000 2000000000 3000000000 4000000000 5000000000 6000000000 7000000000

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 29

GDP in 1000 US$ (year 2002)

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 30

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 31

0 20 40 60 80 100 120 140 160 180 200

Height (in cms) Weight (in kgs)

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 32

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 34

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 36

Mathematics Physics Chemistry Biology

IME602:Probability & Statistics R.N.Sengupta,DoMS.,IIT Kanpur,INDIA 38

Rice Wheat Vegetables Cereals Percentage