Download as pdf or txt
Download as pdf or txt
You are on page 1of 56

u

Zh
Chapter 4

g
an
Continuous Random

qi
ng
Variables and Probability
Ho
Distributions
T
UP
NJ
u
Zh
Section 4.1

g
an
Probability Density

qi
ng
Functions
Ho
T
UP
NJ
Continuous rv’s

u
Zh
A random variable X is continuous if
 Possible values comprise either a single interval on the

g
an
number line (for some A < B, any number x between A
and B is a possible value) or a union of disjoint

qi
intervals

ng
 P(X = c) = 0 for any number c that is a possible value of
Ho
X.
T
UP
NJ
Definition of pdf

u
Zh
 Let X be a continuous rv.
 Then a probability distribution or probability density

g
an
function (pdf) of X is a function f (x) such that for any
two numbers a and b with a ≤ b,

qi
𝑏
𝑃 𝑎 ≤ 𝑋 ≤ 𝑏 = න 𝑓 𝑥 𝑑𝑥
ng 𝑎
Ho
T
UP
NJ
Definition of pdf (cont.)

u
Zh
 That is, the probability that X takes on a value in the
interval [a, b] is the area above this interval and under

g
the graph of the density function.

an
 The graph of f (x) is often referred to as the density

qi
curve.

ng
Ho
T
UP
NJ
Properties/requirements of pdf

u
Zh
For f(x) to be a legitimate pdf, it must satisfy the following
two conditions:

g
an
 f(x) ≥ 0 for all x

qi
 ng
Ho
T
UP
NJ
P(X=c)=0 for a Continuous rv

u
Zh
 When X is a discrete random variable, each possible
value is assigned positive probability.

g
an
 This is not true of a continuous random variable (that
is, the second condition of the definition is satisfied)

qi
because the area under a density curve that lies above

ng
any single value is zero:
Ho
T

 𝑃 𝑎≤𝑋≤𝑏 =𝑃 𝑎<𝑋<𝑏 =𝑃 𝑎<𝑋≤𝑏 =


UP

𝑃(𝑎 ≤ 𝑋 < 𝑏)
NJ
Example: Incompletely Specified pdf

u
Zh
Suppose that the pdf of a certain random variable X has
the following form:

g
cx, 0  x  4,

an
f ( x)  

qi
 0, otherwise,

ng
where c is a given constant. Determine the value of c.
Ho
Solution:
T
UP
NJ

Therefore, c= 1/8;
Example.
Calculating Probabilities from a pdf

u
Zh
Suppose that the pdf of X is

g
 x / 8, 0  x  4,

an
f ( x)  

qi
 0, otherwise,

ng
Determine the values of 𝑃(1 ≤ 𝑋 ≤ 2) and 𝑃(𝑋 > 2).
Ho
2 3
Solutions: P(1  X  2)   1
xdx 
1 8
T

16
UP

4 3
P( X  2)   1
xdx 
NJ

2 8 4
Example

u
Zh
 Suppose that the pdf of a certain random variable X
has the following form:

g
an
qi
ng
Ho
T
UP
NJ
NJ
UP
T
Example (cont.)

Ho
ng
qi
an
g
Zh
u
Example (cont.)

u
Zh
g
an
qi
ng
 Because whenever 0 ≤ 𝑎 ≤ 𝑏 ≤ 360 in this example,
P(a ≤ X ≤ b) depends only on the width b − a of the
Ho
interval, X is said to have a uniform distribution.
T
UP
NJ
Uniform Distribution

u
Zh
 A continuous rv X is said to have a uniform
distribution on the interval [A, B] if the pdf of X is

g
an
qi
ng
Ho
T
UP
NJ
NJ
UP
T
Ho
Homework 4.1(1)

ng
qi
an
g
Zh
u
NJ
UP
T
Ho
Homework 4.1(2)

ng
qi
an
g
Zh
u
u
Zh
Section 4.2

g
an
Cumulative Distribution

qi
ng
Functions and Expected
Ho
Values
T
UP
NJ
Definition of cdf
 The cumulative distribution function F(x) for a

u
Zh
continuous rv X is defined for every number x by

g
an
 For each x, F(x) is the area under the density curve to

qi
the left of x.

ng
Ho
T
UP
NJ
Using F(x) to Compute Probabilities

u
Zh
Let X be a continuous rv with pdf f (x) and cdf F(x). Then
for any number a,

g
𝑃 𝑋 > 𝑎 = 𝑃 𝑋 ≥ 𝑎 = 1 − 𝐹(𝑎)

an
and for any two numbers a and b with a < b,

qi
𝑃 𝑎≤𝑋≤𝑏 =𝑃 𝑎<𝑋<𝑏

ng
=𝑃 𝑎<𝑋≤𝑏 =𝑃 𝑎≤𝑋<𝑏
= 𝐹 𝑏 − 𝐹(𝑎)
Ho
T
UP
NJ
Example

u
Zh
 Suppose the cdf of the magnitude X of a dynamic load
on a bridge (in newtons) is given by

g
an
qi
ng
 Then the probability that the load is between 1 and
Ho
1.5 is
T
UP
NJ
Example (cont.)

u
Zh
 The probability that the load exceeds 1 is

g
an
qi
ng
 In-class exercise: The probability that the load is
between 1 and 2 is
Ho
0.688.
T
UP
NJ
Obtaining f (x) from F(x)

u
Zh
 If X is a continuous rv with pdf f(x) and cdf F(x), then
at every x at which the derivative F’(x) exists,

g
an
F’(x) = f(x).

qi
ng
Ho
T
UP
NJ
Example: Calculating a pdf from a cdf

u
Zh
Let the cdf of a random variable be

g
 0, x  2

an
 1 3  x3 
F ( x)     4 x  ,  2  x  2

qi
 2 32  3
 x2
ng
Ho 1,
Find the pdf.
T

Solution:
UP

3
 (4  x 2 ),  2  x  2
NJ

f ( x)  F ' ( x)   32
 0, otherwise
Expected Values

u
Zh
 The expected or mean value of a continuous rv X with
pdf f (x) is

g
an
qi
 If X is a continuous rv with pdf f(x) and h(X) is any
function of X, then
ng
Ho
T
UP
NJ
Example

u
Zh
The pdf of weekly gravel sales X was

g
an
qi
ng
So Ho
T
UP
NJ
In-class Exercise

u
Zh
Compute the expected value of a uniform distribution
with pdf

g
 1

an
 a  x  b,
f ( x)   b  a

qi
0 otherwise.

Solution: ng
Ho

ab
b
x
E ( X )   xf ( x)dx   dx 
T

ba 2
UP

 a
NJ
Example

u
Zh
Let V be a continuous random variable with pdf
1

g
 0  v  a,

an
f (v )   a

qi
0 otherwise.

ng
Compute the expected value of 𝑊 = 𝑘𝑉 2 .
Ho
Solution:
T

 a
UP

1 1 2
E (W )   kv f (v)dv   kv dv  ka .
2 2

a 3
NJ

 0
In-class Exercise

u
Zh
Let X be a continuous random variable with pdf

g
e  x x  0,

an
f ( x)  
0 x  0.

qi
ng
Compute the expected value of 𝑌 = 𝑒 −2𝑋 .
Ho
Solution:
T

 
UP

1
E (Y )   e 2 x
f ( x)dx   e 2 x  x
e dx  .
NJ

 0
3
Variance of a Continuous rv

u
Zh
 The variance of a continuous random variable X with
pdf f(x) and mean value 𝜇 is

g
an
qi
ng
 The standard deviation (SD) of X is
Ho
T
UP
NJ
A Shortcut Formula for V(X)

u
Zh
V(X)=E(X2)-[E(X)]2

g
an
Proof:V(X)=E[X-E(X)]2

qi
=E{X2-2XE(X)+[E(X)]2}

ng
Ho
=E(X2)-2[E(X)]2+[E(X)]2

=E(X2)-[E(X)]2
T
UP
NJ

We always use this formula to compute the variance !


Example

u
Zh
The pdf of weekly gravel sales X was

g
an
qi
ng
We have computed Ho
T
UP
NJ
NJ
UP
T
Example (cont.)

Ho
ng
qi
an
g
Zh
u
In-class Exercise

u
Compute the variance of a uniform distribution with pdf

Zh
 1
 a  x  b,
f ( x)   b  a

g
0 otherwise.

an

ab
b
x

qi
Solution: We computed E ( X )   xf ( x)dx   dx  .
 a
ba 2

ng
So V ( X )  E ( X 2 )  E ( X ) 2
Ho
x
b
1 ab
dx  
b  a
2 2

 
2
.
T

ba  2  12
UP

b  a
NJ

ab
2
Conclusion: E( X )  , V (X ) 
2 12
Homework 4.2(1)

u
Zh
The cdf of an rv X is

g
an
qi
Use this to compute the following:

ng
a. The probability density function f(x)
b. P(X ≤ 1)
Ho
c. P(0.5 ≤ X ≤ 1)
T

d. P(X > 0.5)


UP

e. E(X)
NJ

f. V(X) and 𝜎𝑋
u
Zh
Section 4.3

g
an
The Normal Distribution

qi
ng
Ho
T
UP
NJ
Motivation

u
Zh
 The normal distribution is the most important one in
all of probability and statistics.

g
an
qi
 Before defining it, let’s see why it is so important with
a simple but informative example…

ng
Ho
 Suppose you have n dice. You roll them and count the
total number of dots you see. Call this random
T

variable X. What does it look like?


UP
NJ
NJ
UP
T
Ho
ng
qi
an
g
Zh
u
Motivation (cont.)

u
Zh
 These probability mass functions look more and more
like a bell-shaped curve…

g
an
qi
 Remarkably the sum of many independent random
variables almost always looks similar to a bell curve !!!

ng
No matter what these random variables look like
Ho
themselves…
T

 This bell curve is known as the normal distribution.


UP
NJ
Definition and Properties

u
 A continuous rv X is said to have a normal distribution

Zh
with parameters 𝜇 and 𝜎 (or 𝜇 and 𝜎 2 ), where −∞ <
𝜇 < ∞ and 0 < 𝜎, if the pdf of X is

g
( x )2
1 

an
f ( x;  ,  )  e ,   x  
2 2

2 

qi
 The statement that X is normally distributed with

ng
parameters 𝜇 and 𝜎 2 is often abbreviated 𝑋~𝑁(𝜇, 𝜎 2 ).
 Clearly 𝑓(𝑥; 𝜇, 𝜎) ≥ 0 but a somewhat complicated
Ho
calculus argument must be used to verify that

‫׬‬−∞ 𝑓 𝑥; 𝜇, 𝜎 𝑑𝑥 = 1.
T
UP

 It will be proved later that 𝐸 𝑋 = 𝜇 and 𝑉 𝑋 = 𝜎 2 ,


NJ

so the parameters are the mean and variance of X.


Shapes of the Normal Distributions

u
 The density curve is symmetric about 𝜇 and bell-

Zh
shaped, so the center of the bell (point of symmetry)
is the mean of the distribution.

g
an
P  μ  h  X  μ

qi
 P  μ  X  μ  h

ng
Ho
T
UP
NJ
Shapes with different μ’s

u
Zh
f(x)

g
 =5

an
 =5

qi
ng
Ho

 1
T

O x
UP
NJ
Shapes with different σ’s

u
Zh
f(x)
 0.5
0.798

g
an
 1

qi
0.399  1.5
0.266 ng
Ho
T


UP

O x
NJ
Why Standard Normal Distribution?

u
Zh
 To compute 𝑃(𝑎 ≤ 𝑋 ≤ 𝑏) when X is a normal rv with
parameters 𝜇 and 𝜎, we must determine

g
an
qi
 Unfortunately, none of the standard integration
ng
techniques can be used to evaluate this expression.
Ho
 Instead, for 𝜇 = 0 and 𝜎 = 1, the expression has been
calculated using numerical techniques and tabulated
T

for certain values of a and b.


UP

 This table can also be used to compute probabilities


NJ

for any other values of 𝜇 and 𝜎 under consideration.


Definition
 The normal distribution with parameter values 𝜇 = 0

u
Zh
and 𝜎 = 1 is called the standard normal distribution.
 A random variable having a standard normal

g
distribution is called a standard normal random

an
variable and will be denoted by Z.

qi
 The pdf of Z is

ng
Ho
 The graph of 𝑓(𝑧; 0,1) is called the standard normal
T

(or z) curve.
UP

𝑧
 The cdf of Z is 𝑃 𝑍 ≤ 𝑧 = ‫׬‬−∞ 𝑓(𝑦; 0,1) 𝑑𝑦, which
NJ

we will denote by Φ(𝑧).


Shapes and properties

u
Zh
 (x ) (x )

g
an
qi
ng
Ho
T

1
1 0  ;
UP

2
NJ

2 x  R ,  x   1   x  ;
Standard Normal Table
 The standard normal distribution does not frequently

u
Zh
serve as a model for a naturally arising population.
 Instead, it is a reference distribution from which

g
information about other normal distributions can be

an
obtained.

qi
 Appendix Table A.3 (page 672) gives Φ 𝑧 = 𝑃(𝑍 ≤ 𝑧),
the area under the standard normal density curve to
ng
the left of z, for z = -3.49, -3.48, . . . , 3.48, 3.49.
Ho
T
UP
NJ
Example

u
Zh
 Let’s determine the following standard normal
probabilities.

g
an
 𝑃(𝑍 ≤ 1.25)
 Solution: 𝑃 𝑍 ≤ 1.25 = Φ 1.25 = 0.8944

qi
 𝑃(𝑍 > 1.25)
ng
 Solution: 𝑃 𝑍 > 1.25 = 1 − Φ 1.25 = 0.1056
Ho
T
UP
NJ
Example (cont.)

u
 𝑃(−0.38 ≤ 𝑍 ≤ 1.25)

Zh
 Solution: 𝑃 −0.38 ≤ 𝑍 ≤ 1.25 = Φ 1.25 −
Φ −0.38 = 0.8944 − 0.3520 = 0.5424

g
an
qi
ng
Ho
 Exercise: 𝑃(𝑍 ≤ −1.25)
T
UP

 Solution: 𝑃 𝑍 ≤ −1.25 = Φ −1.25 = 1 −


Φ 1.25 = 1 − 0.8944 = 0.1056
NJ
Relationship between a standard and
nonstandard normal rv

u
Zh
g
an
qi
ng
Ho
T
UP
NJ
Proof
NJ
UP
T
Ho
ng
qi
an
g
Zh
u
Example

u
Zh
g
an
qi
ng
Ho
T
UP
NJ

P(X<10)=? P(2<X<10)=? P(X<0)=?


The Empirical Rule

u
Zh
If the population distribution of a variable is
(approximately) normal, then

g
an
1. Roughly 68% of the values are within 1 SD of the
mean.

qi
2. Roughly 95% of the values are within 2 SDs of the
mean.
ng
Ho
3. Roughly 99.7% of the values are within 3 SDs of the
mean.
T
UP

In statistics, the empirical rule is also known as the 68–


NJ

95–99.7 rule or three-sigma rule.


Graph of the Empirical Rule

u
Zh
g
an
qi
ng
Ho
3 2   2 3
68.26%
T
UP

95.44%
NJ

99.74%
Example: Verify the Empirical Rule

u
Zh
P(| X   |  )  (1)  (1)  0.6826

g
P(| X   | 2 )  (2)  (2)  0.9544

an
P(| X   | 3 )  (3)  (3)  0.9974

qi
ng
Ho
T
UP
NJ
Expected Value and Variance of Standard
Normal Distribution

u
Zh
x2
1 
If 𝑋~𝑁 0,1 , then  ( x)  e 2
   x  
2

g
an
  x2
1 
E ( X )   x ( x)dx   xe 2
dx  0
2

qi
 

ng
  x2
1 
E ( X )   x  ( x)dx 
2 2
x e
2 2
dx  1
2
Ho
 

V ( X )  E ( X 2 )  ( E ( X )) 2  1  0  1
T
UP

Conclusion:
E ( X )  0, V ( X )  1
NJ
Expected Value and Variance of Normal
Distribution

u
Zh
𝑋−𝜇
If 𝑋~𝑁 𝜇, 𝜎 2 , then 𝑍 = ~𝑁(0,1). We have just
𝜎
proved that

g
E ( Z )  0, V ( Z )  1

an
qi
Since X  Z   ,

ng
E ( X )  E (Z   )  E ( Z )    
Ho
V ( X )  V (Z   )   2V ( Z )  0   2
T
UP

Conclusion: E ( X )   ,V ( X )   2
NJ
Homework 4.3(1)

u
Zh
Let 𝑋~𝑁(2, 𝜎 2 ),
a) Compute 𝑃(𝑋 > 2).

g
an
b) If 𝑃 0 < 𝑋 < 2 = 0.4, compute 𝑃 𝑋 > 4 .
c) If 𝜎 = 4, using the table of Standard Normal CDF at

qi
the end of the textbook and compute the following

ng
probabilities:
𝑃 𝑋≤6
Ho
𝑃 𝑋 > −3
𝑃(−5 ≤ 𝑋 < 1)
T

𝑃(|𝑋| < 2)
UP

Note: the table can also be found in the webpage


NJ

https://en.wikipedia.org/wiki/Standard_normal_table

You might also like