Advanced Probability Models

AQ088-3-2-APROM & VD 1

Univariate Distributions
 Topic & Structure of the lesson
Topic Outline
 Introduction
 Discrete Uniform Distribution
 Binomial Distribution
 Poisson Distribution
 Geometric Distribution
 Negative Binomial Distribution
 Hypergeometric Distribution
 Multinomial Distribution

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 2 (of 99)

 Topic & Structure of the lesson
Topic Outline
 Continuous Uniform Distribution
 Normal Distribution
 Exponential Distribution
 Gamma and Pareto Distribution
 Lognormal, Beta, Weibull and Chi-
Square Distribution

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 3 (of 99)

 Learning Outcomes

On completion of this chapter

you
should be able to:
• Know the theoretical aspect of discrete
distribution
• Know the theoretical aspect of continuous
distribution

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 4 (of 99)

 Key Terms You Must Be
Able To Use
• If you have mastered this topic, you should be able
to use the following terms correctly in your
assignments and exams:

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 5 (of 99)

 Discrete Probability Distribution
 If a random variable is a discrete variable, its
probability distribution is called a discrete
probability distribution.

 A random varibale is discrete if it has finite or

countable number of possible outcomes that can
be listed.

 A discrete probability distribution lists each

possible value the random variable can assume,
together with its probability.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 6 (of 99)

 Discrete Probability Distribution
 A discrete probability distribution must satisfy
the following conditions:
i. The probability of each value of the discrete
random variable is between 0 and 1.

0 P ( X ) 1
ii. The sum of all the probabilities is 1.

 P( X ) 1

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 7 (of 99)

 Discrete Probability Distribution

Type of distribution:

 Uniform Discrete
 Hypergeometric
 Bernoulli
 Binomial
 Poisson
 Geometric
 Negative Binomial

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 8 (of 99)

 Discrete Uniform Distribution
 If a random variable can take on n different
values with equal probability, it has a
discrete uniform distribution.
 A random variable X has a uniform discrete
distribution and it is referred to as a
discrete uniform random variable if and
only if its probability distribution is given by
1
f ( x)  a  x b
b  (a  1)

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 9 (of 99)

 Uniform Discrete Distribution
 The CDF for uniform discrete distribution
 0 xa
 int( x)  (a  1)
F ( x)  a x  b
 b  (a  1)
 1 x b
 If
1  x n
1
f ( x)  1  x n
n

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 10 (of 99)

 Uniform Discrete Distribution
 The mean and variance for uniform
discrete distribution
1 b
b(b  1)  a(a  1)
  x[ ]
a b  (a  1) 2(b  (a  1))
2
(b  a  1)  1
 E ( X )  [ E ( X )] 
2 2 2

12
 If
1  x n
n 1 n2  1 n2  1
  
2 
2 12 12

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 11 (of 99)

 Uniform Discrete Distribution
 The median and MGF for uniform discrete
distribution
a b
median 
2
e at  e (b 1)t
MGF 
[b  (a  1)](1  et )
 If
1  x n
( n 1) t t
n 1 e e
median  MGF 
2 n(et  1)

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 12 (of 99)

 Uniform Discrete Distribution
Example for a Uniform Discrete pmf

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 13 (of 99)

 Bernoulli Distribution
 Variable X is a Bernoulli random variable
if it takes on only two values 0 and 1 and
its PMF and CDF is
q 1  p x 0

f ( x)  p x 1
 0 otherwise


 0 x0

F ( x) 1  p 0  x  1
 1 x 1

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 14 (of 99)
 Bernoulli Distribution
 The mean, median, variance and MGF for
Bernoulli distribution
 p   pq
2

 0 pq

median 0.5 p q
 1 pq

t
MGF q  pe

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 15 (of 99)

 Bernoulli Distribution
Example for a Bernoulli pmf

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 16 (of 99)

 Geometric Distribution
 In Bernoulli trials until a 1 (success) occurs.
 If the probability of a 1 on any single trial is
p, the probability that the first success will
occur on the k-th trial is p(1-p)k-1.
 The PMF and CDF for a Geometric random
variable with k attempts is

F(x) = 1 – (1 – p)k

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 17 (of 99)

 Geometric Distribution
 The mean, variance and MGF for
Geometric distribution
1 q
   2 2

p p

pet
MGF  t
1  qe

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 18 (of 99)

 Geometric Distribution
Example for a Geometric pmf

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 19 (of 99)

 Geometric Distribution
 Some texts use a different version of this
distribution, which only cout failures until the first
success. The two distributions differ by a constant
value of one.
 The PMF and CDF for this form of random variable
with k failures is

 p(1  p) k 0 k 
f ( x) 
 0 otherwise

F ( x) 1  (1  p) k 1
AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 20 (of 99)
 Geometric Distribution
 The mean, variance and MGF for this
distribution
1 p q
   2 2

p p

p
MGF 
1  qet

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 21 (of 99)

 Binomial Distribution
 In Bernoulli trials if X represents the total
number of 1’s (successes) that occur in those n
trails, then X is said to be a Binomial random
variable.
 The PMF for a Binomial random variable with k
successes in n attempts is

 n  k n k
  p (1  p) 0 k n
f ( x)  k 
 0 otherwise


AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 22 (of 99)

 Binomial Distribution
 The mean, variance and MGF for Binomial
distribution

 np

 npq np(1  p)
2

MGF (q  pet ) n

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 23 (of 99)

 Binomial Distribution
Example for a Binomial pmf

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 24 (of 99)

 Negative Binomial Distribution
 In Bernoulli trials until k-th 1 (success)
occurs on n-th trial.
 The PMF for a Negative Binomial or Pascal
random variable with k successes in n
attempts is

 n  1 k
  p (1  p) n  k 0 k n
f ( x)  k  1
 0 otherwise


AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 25 (of 99)

 Negative Binomial Distribution
 The mean, variance and MGF for
Negative Binomial distribution
k (1  p)

p
k (1  p)
 
2

p2
k
 p 
MGF  
t 
 1  qe 

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 26 (of 99)

 Negative Binomial Distribution
Example for a Negative Binomial pmf

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 27 (of 99)

 Quick Review Question

Each day a bakery delivers the same number of

loaves of bread to a certain shop which sells, on
average, 98% of them. Assuming that the number
of loaves sold per day has a binomial distribution
with a standard deviation of 7, find the number of
loaves of bread the shop would expect to sell per
day.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 28 (of 99)

 Quick Review Question

Let X be the random outcome of tossing a fair

die, and Y1= min(X1,X2,X3), Y2 (where Y1≤Y2 ≤Y3)
and Y3= max(X1,X2,X3) be the order statistics from
a random sample X1, X2, X3 from the distribution
of X. Find P(Y1=3)

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 29 (of 99)

 Hypergeometric Distribution
 The hypergeometric distribution is a discrete
probability that describes the probability of k
successes in n draws, without replacement, from
a finite population of size N that contains exactly
K successes.
 The PMF for a Hypergeometric Distribution is

 K  N  K 
   
 k  n  k 
f ( x) 
N
 
n
AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 30 (of 99)
 Hypergeometric Distribution
 The mean and variance for
Hypergeometric distribution
K
 n
N

nK ( N  K )( N  n)
  2

N 2 ( N  1)

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 31 (of 99)

 Poisson Distribution
 This distribution often used to model the number of
incidences of some characteristic in time or space:
 Arrivals of customers in a queue
 Number of animals in an area
 In general, the Poisson distribution will provide a good
approximation to binomial probabilities when n ≥ 30 and p ≤
0.05. When n ≥ 100 and np <10, the approximation will
generally be excellent.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 32 (of 99)

 Poisson Distribution
 The PMF, mean, variance and MGF for
Poisson distribution

e   x
f ( x)  x 0,1,2,...
x!
   2 

 ( e t  1)
MGF e

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 33 (of 99)

 Quick Review Question
N is a Poisson random variable such that

P ( N 1) 2 P ( N 2)

Find the variance of N.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 34 (of 99)

 Multinomial Distribution
 The multinomial distribution is a generalization of the
binomial distribution. The multinomial distribution
arises from an extension of the binomial experiment
to situations where each trial has k ≥ 2.
 The PMF for a Multinomial Distribution is

n! x x x
f ( x)  p1 1 p2 2 .... pk k
x1! x2 !...xk !

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 35 (of 99)

 Multinomial Distribution
 The mean and variance for Multinomial
distribution

E ( X i ) npi

Var ( X i ) npi (1  pi )

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 36 (of 99)

 Quick Review Question
John rolls 8 fair dice. Find the probability that he
obtains
(a) exactly two 3s.
(b) exactly two 1s, two 2s, two 3s and two 4s.
(c) more than one 5 and more than one 6.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 37 (of 99)

 Continuous Probability Distribution
 If a random variable is a continuous variable, its
probability distribution is called a continuous
probability distribution.
A continuous probability distribution is a
probability distribution in which the random
variable X can take on any value.
 Because there are infinite values that X could
assume, the probability of X taking on any one
specific value is zero. Therefore we often speak in
ranges of values.
E.g. P(X>2) = 0.3

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 38 (of 99)

 Continuous Probability Distribution
 The equation used to describe a continuous
distribution is called a probability density function
(PDF).
 All PDF satisfy the following conditions:
• The random variable Y is a function of X; that is y =
f(x)
• The value of y is greater than or equal to zero for
all values of x.
• The total area under the curve of the function is
equal to one.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 39 (of 99)

 Continuous Probability Distribution
 A PDF given over a range a ≤ x ≤ b satisfies the
following:
• f(x) ≥ 0 for all the values of x lying between a
and b.
 The probability can be found out by integrating
the density function with respect to x over the
interval [a, b]

b
P (a  X b)  f ( x)dx
a

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 40 (of 99)

 Quick Review Question

A continuous random variable X has pdf f(x)

where 2
k ( x  2)  2 x  0
 4
f ( x)  4k 0 x 
 3
 0 otherwise
Find
(a) P(X< -1)
(b) P(X>1)
(c) P(|X| ≤ 0.5)

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 41 (of 99)

 Quick Review Question

The lifetime of a machine part has a continuous

distribution on the interval (0, b) with probability
density function f, where f(x) is proportional to
(10+x)-2. Find P(X>6) if the median of the lifetime
of the machine is 5.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 42 (of 99)

 Quick Review Question

The loss amount, X, for an insurance policy has

cumulative distribution function
 0 x0
3
1 x
F ( x)  (2 x 2  ) 0  x  3
9 3
 1 x 3

Calculate the mode of the distribution.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 43 (of 99)

 Continuous Probability Distribution

Type of distribution:

 Uniform Distribution
 Normal Distribution
 Exponential Distribution
 Gamma and Pareto Distribution
 Lognormal, Beta, Weibull and Chi-Square
Distribution

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 44 (of 99)

 Uniform Distribution
 A uniform distribution is one for which the
probability of occurrence is the same for all values of
X. It is sometimes called a rectangular distribution.

 A uniform random variable on the interval [a, b] is

the random variable that chooses randomly and
equally likely a point in [a, b]. E.g. A bus may arrive
anytime equally between 10 am and 11 am. Let’s X
denotes the time at which the bus arrives. Then
X~U(0,60), where the measure unit is minutes.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 45 (of 99)

 Uniform Distribution
 The PDF and CDF for a continuous uniform
distribution on the interval [a, b] are

 0 xa
 1
f ( x)  a  x b
b  a
 0 x b

 0 xa
x  a
F ( x)  a x  b
b  a
 1 x b

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 46 (of 99)

 Uniform Distribution
 In a uniform distribution, the value of mean is
same as the value of median where

1
 median  (a  b)
2
 The variance for uniform distribution

2
(b  a)
 
2

12

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 47 (of 99)

 Uniform Distribution
 The MGF for uniform distribution

 etb  eta
 for t 0
MGF  t (b  a)
 1 for t 0

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 48 (of 99)

 Uniform Distribution
The PDF for a Uniform Distribution

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 49 (of 99)

 Uniform Distribution
The CDF for a Uniform Distribution

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 50 (of 99)

 Quick Review Question
John spins a ‘Wheel of Fortune’ at a fair, where
the radius of the wheel is 10 unit. When the wheel
stops, the shortest distance of an arrow
measured along the circumference from John is
denoted by C.
Find
(a) P(C>12).
(b) E(C3)

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 51 (of 99)

 Quick Review Question

Two people agree to meet each other on a

particular day, between 5 and 6 PM, They arrive
independently on a uniform time between 5 and 6
and wait for 15 minutes. What is the probability
that they meet each other?

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 52 (of 99)

 Quick Review Question

Ram and Shyam wanted to meet at a park about

12.30 P.M.. Given that Ram arrives at a time
uniformly distributed between 12.15 P.M. to 12.45
P.M. and Shyam independently arrives at a time
uniformly distributed between 12.00 P.M. and
1.00 P.M.. Find the probability that the first to
arrive waits no longer than 5 minutes.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 53 (of 99)

 Normal Distribution
 The normal distribution or Gaussian distribution is a continuous
probability distribution that symmetric and have bell-shaped
density curve with a single peak.
 The normal distribution is useful because of the central limit
theorem.
 The normal distribution is symmetric and centered on the mean
(same as the median and mode). While the x-axis ranges from
negative infinity to positive infinity.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 54 (of 99)

 Normal Distribution
 Nearly all the X values fall within +/- three standard
deviations of the mean (99.7% of the values are fall
within µ±3σ), while about 68% are within +/- one
standard deviation and about 95% are within +/- two
standard deviations. This is often called the three
sigma rule or the 68-95-99.7 rule.
 The pdf for a normal distribution is

 ( x  )2
1 2 2
f ( x)  e
 2
AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 55 (of 99)
 Normal Distribution
 The standard normal probability function has a mean of
zero and a standard deviation of one. Often times the x
values of the standard normal distribution are called z-
scores.
 We can covert the normal distributions to the standard
normal distribution using the equation below, and then
we can calculate probabilities using a normal distribution
table (z-table)

X
z

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 56 (of 99)
 Quick Review Question

Let X be a normal random variable with mean 0

and variance k > 0. Find P(X2 < k)

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 57 (of 99)

 Quick Review Question

Two stalls, A and B are selling sandwiches. The

sandwiches of stall A have masses normally
distributed with mean 65g and standard deviation
4g. The masses of the sandwiches of stall B are
normally distributed with mean 100g and standard
deviation 5g. Find the probability that
(a)the mass of 3 sandwiches of stall A is greater
than the mass of 2 sandwiches of stall B.
(b)the mean of the mass of 16 sandwiches of stall A
is greater than 66g.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 58 (of 99)

  Normal distribution can be used as an approximation to the
binomial distribution when:
 n is large (greater than 30)
 p is not too small or large (the closer to 0.5 the better)
 when used as an approximation, Normal distribution has

μ np
σ  npq

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 59 (of 99)

  Continuity Correction
 Used when Normal distribution are used for discrete variables
rather than continuous variables.
 When the range of values of the discrete variable is large, the
continuity correction can be ignored because it will be
significant.
 However, it ought to be used when the range is small enough to
make a difference of 0.5 potentially significant.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 60 (of 99)

 Quick Review Question

A man is playing blackjack . His probability of

winning at each game is 0.45, probability of losing
is 0.5, and the probability of breaking even is
0.05. At every game, he either wins $10, loses
$10, or breaks even, and he plays 100 times.
Using normal approximation, estimate the
probability that he actually comes out ahead in
those 100 games.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 61 (of 99)

 Exponential Distribution
 The exponential distribution is one of the widely
used continuous distributions. It is often to used
to model the time elapsed between events.
 The PDF for exponential distribution is

where x
1 
 x
f ( x)  e 
e 0 x  

1



AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 62 (of 99)

 Exponential Distribution
 The mean, median, variance and CDF for
Exponential distribution

1 1
      2 2

 
2

ln 2
median  ln 2 

 0 x0
F ( x)   x
1  e x 0

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 63 (of 99)

 Quick Review Question

Let X have a distribution with the 75-th percentile

1
equal to and density function equal to
3
e  x x 0
f ( x) 
 0 otherwise
What is  ?

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 64 (of 99)

 Quick Review Question

Assume the number of train delays during a year

follows a Poisson distribution with the mean of
10. Find the probability that the time between the
next two delays is at least 3 months.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 65 (of 99)

 Exponential Distribution
 One of the most important properties of the
exponential distribution is the memoryless
property or lack of memory property

P ( X  x  y | X  x ) P ( X  y )

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 66 (of 99)

 Quick Review Question

The time between arrivals of taxis at a busy

intersection is exponentially distributed with a
mean of 10 minutes. What is the probability that
(a) a taxi will arrive in next 10 minutes?
(b) you wait longer than one hour?
(c) one taxi arrives within the next 10 minutes,
given that you have already been waiting for one
hour?

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 67 (of 99)

 Erlang Distribution

 An exponential random variable describes

the length until the first count is obtained in
a Poisson process. A generalization of the
exponential distribution is the length until r
counts occur in a Poisson process has an
Erlang random variable.

 The probability density function for an

Erlang random
r x r  1e  x variable is
f ( x)  , for x  0 and r 1,2,...
(r  1)!

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 68 (of 99)

 Gamma Distribution

 The Erlang distribution is a special case of

the gamma distribution. If the parameter r
of an Erlang variable is not an integer, but r
> 0, the random variable has a gamma
distribution.

 The gamma distribution is frequently a

probability model for waiting times; for
instance, in life testing, the waiting time
until death is a random variable that is
frequently modeled with a gamma
distribution.
AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 69 (of 99)
 Gamma Distribution

 The PDF for Gamma Distribution

x

r x r  1e  x x r  1e 
f ( x)   r , for 0 x  
(r )  (r )
 (r ) is gamma function where

(r )  x r  1e  x dx (r  1)! 1
and
( )  
0
2
 The exponential distribution is a special
case of the gamma distribution with r = 1.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 70 (of 99)

 Gamma Distribution
 The mean, variance moment generating
function for a gamma distribution

r
  r

r
  2 r 2
2


r
  
M X (t )  
 t

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 71 (of 99)

 Example

Suppose the number of customers per hour

arriving at a shop follows a Poisson process
with mean 30. What is the probability that the
shopkeeper will wait more than 5 minutes
before both of the first two customers arrive?

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 72 (of 99)

 Example
If X denotes the waiting time in minutes until the second
customers, then X has a gamma distribution with r = 2, θ =
1/λ = 2.
x x
 2 1 2 
x e xe 2
P ( X  5)  dx  dx
5
2 2 2
5
4
x x 
1 
   2 xe  4e 
2 2
4 5
5
7 2
 e 0.287
2

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 73 (of 99)

 Quick Review Question

Engineers designing the next generation of space

shuttles plan to include two fuel pumps —one
active, the other in reserve. If the primary pump
malfunctions, the second is automatically brought
on line. Suppose a typical mission is expected to
require that fuel be pumped for at most 50 hours.
According to the manufacturer's specif ications,
pumps are expected to fail once every 100 hours.
What are the chances that such a fuel pump
system would not remain functioning for the full 50
hours?

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 74 (of 99)

 Quick Review Question

The time between arrivals of customers at an

ATM is an exponential random variable with a
mean of 5 minutes. What is the probability that
(a)More than three customers arrive in 10
minutes?
(b)The time until the fifth customer arrives is less
than 15 minutes?

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 75 (of 99)

 Chi-square Distribution
 It is one of the most widely used probability
distribution in inferential statistics, e.g., in hypothesis
testing or in construction of confidence intervals.

 Chi-square distribution is a special case of the gamma

distribution where r = n/2 and λ = ½, where n is a
positive integer. The PDF for Chi-square distribution

n x
1 1
f ( x)  n
x 2
e 2
, for 0  x  
 n  2 distribution with n degrees
is called the chi-square
of freedom   2
 2

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 76 (of 99)

 Chi-square Distribution
 The mean, variance and moment generating
function for a chi-square distribution

 n

 2n
2

n
 1  2
M X (t )  
 1  2t 

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 77 (of 99)

 Chi-square Distribution
 Chi-square distribution with n degrees of freedom
is the distribution of a sum of the squares of n
independent standard normal variables.

 If Z1, Z2,...Zn is a random sample from a standard

normal distribution, then Z12+ Z22+...Zn2 has the
chi-square distribution with n degrees of
freedom. If X and Y are independent and have the
standard normal distribution then X2 + Y2 has the
chi-square distribution with 2 degrees of
freedom. X2 + Y2 is also has the exponential
distribution with mean 2 (or hazard rate λ = 0.5)

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 78 (of 99)

 Chi-square Distribution
 Chi-square distribution is used in
hypothesis testing when it is
wished to compare an actual,
observed distribution with a
hypothesized, or expected
distribution.
O  E 2

 2 = E
 where O = the observed frequency of any
value E = the expected frequency
of any value

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 79 (of 99)

  The obtained value from the formula is
compared with the value from 2 table for
a given significance level and the number
of degrees of freedom.
 Degrees of freedom = (Rows-1)(Columns –
1)

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 80 (of 99)

 Pareto Distribution

 The Pareto distribution, named after the Italian civil

engineer, economist and sociologist Vilfredo Pareto, is a
power law probability distribution that is used in
description of social, acturial and many other types of
observation (e.g. Pareto distributions can be used to
model the distribution of full time salaries, especially
when there is a minimum wage).

 Pareto distribution are clearly right skewed.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 81 (of 99)

 Pareto Distribution
 A Pareto distribution is defined by two parameters xm, the
minimum value the random variable can take (i.e. The
minimum wage) and α, where xm > 0 and α > 0.

 The parameter α describes the degree of concentration

of the distribution. The smaller α (i.e. the closer α is to 0),
the heavier the tail of the distribution (i.e. the proportion
of individuals earning very high salaries increases).

 The PDF of a Pareto distribution is given by

  xm
  1 x  xm
f ( x)  x

 0 otherwise
AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 82 (of 99)
 Pareto Distribution
 For α > 1
 xm 
E( X )  
1 1
 In order to calculate probabilities for the Pareto
distribution, we use
 1, x  xm
 
P ( X  x)  xm 
 x  , x  xm

together with the rules
P ( a  X  b) P ( X  a )  P ( X  b)
P ( X  x) 1  P ( X  x)
AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 83 (of 99)
 Example
Suppose that monthly salaries, denoted X,
have a Pareto (1000, 2) distribution. Calculate
(a) the expected salary
(b) the probability that an individual earns
(i) above the mean salary.
(ii) below1500
(iii) between 3000 and 6000
(c) the median salary

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 84 (of 99)

 Example
xm 2 1000
(a) E( X )   2000
1 2 1

(b)  2
 xm   1000 
(i) P(X>2000) =    0.25
 2000   2000 
 2
 xm   2 5
(ii)P(X<1500) = 1 – P(X>1500)1    1    
 1500   3 9

(iii)P(3000<X<6000) = P(X>3000) –
P(X>6000) 2 2
 1000   1000  1 1 1
      
 3000   6000  9 36 12

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 85 (of 99)

 Example
(c)Let m be the median
P(X>m) = 0.5
2
 1000 
  0.5
 m 
1000
m 1414
0 .5

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 86 (of 99)

 Weibull Distribution
 A Weibull distribution is often used to model the time until
failure of many different physical systems.

 The PDF and CDF of a Weibull distribution with scale

parameter α>0 and shape parameter β>0 is given by


1  x
 x  
 
f ( x)    e , x 0
  
 x 
 

F ( x) 1  e 

 It is seen that when β = 1, the Weibull distribution is identical

to the exponential distribution.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 87 (of 99)

 Weibull Distribution
 The mean, variance and moment function for a
Weibull distribution

 1
 E (X )  1  
 
k  k
E ( X )   1  
k

 
  2    1  2 
 2  2   1      1    
         

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 88 (of 99)

 Example
The time of failure (in hours) of a bearing in a
mechanical shaft is satisfactorily modeled as
a Weibull random vaiable with β = 0.5 and α =
5000 hours. Determine
(a)the mean time until failure.
(b)the probability that a bearing lasts at least
6000 hours.

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 89 (of 99)

 Example

  1 
(a) E ( X ) 5000 1     5000(3) 10000
  0.5  
(b)P(X>6000)
= 1 – F(6000)
  
0.5
 6000 


1  1  e  5000  
e  1.2 0.5
0.334
 
 

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 90 (of 99)

 Beta Distribution
 A Beta distribution with r and s are positive parameter
has the PDF

( r  s ) r  1
f ( x)  x (1  x) s  1 , 0  x 1
( r ) ( s )

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 91 (of 99)

 Beta Distribution
 The mean, variance and moment function for a
Beta distribution

r
 E ( X ) 
r s
k (r  s)(r  k )
E( X ) 
(r )(r  s  k )
rs
 
2

(r  s ) 2 (r  s  1)

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 92 (of 99)

 Lognormal Distribution
 We say that X has the lognormal distribution if its natural
logarithm has the normal distribution. The lognormal
distribution is usually specified by the parameters of its
corresponding normal distribution, i.e., X is lognormal
with parameter μ and σ2 if W = ln X ~ N(μ, σ2).

 The PDF for lognormal distribution

(ln x   ) 2
1 
2 2
f ( x)  e , x 0
x 2

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 93 (of 99)

 Lognormal Distribution
 The mean and variance for a Lognormal
distribution

1
  2
 E ( X ) e 2

 2
  e  1e
2
 2   2

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 94 (of 99)

 Quick Review Question

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 95 (of 99)

 Summary of Main Teaching Points

• Discrete Uniform Distribution

• Binomial Distribution
• Poisson Distribution
• Geometric Distribution
• Negative Binomial Distribution
• Hypergeometric Distribution
• Multinomial Distribution

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 96 (of 99)

 Summary of Main Teaching Points

• Continuous Uniform Distribution

• Normal Distribution
• Exponential Distribution
• Gamma and Pareto Distribution
• Lognormal, Beta, Weibull and Chi-
Square Distribution

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 97 (of 99)

 Question and Answer Session

Q&A

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 98 (of 99)

 Next Session

Joint, Marginal, and

Conditional Distributions

AQ017-3-1-APROM Advanced Probability Models Univariate Distributions Slide 99 (of 99)