05 ES Random Variables and Their Probability Distributions

05 ES Random Variables and their Probability Distributions (1)
Random Variables and their Probability Distributions

A variable, such as the strength of a concrete or any other material or physical quantity, whose value
is uncertain or unpredictable or nondeterministic is called a random variable or a variate if its distribution is
known. A random variable may assume some value, the magnitude of which depends on an outcome of an
experiment in which tests are made and records maintained. A random variable can be formally viewed as a
function defined on the sample space of an experiment such that there is a numerical value of the random
variable corresponding to each possible outcome; that is, there is a probability associated with each
occurrence in the sample space. In contrast to a simple event, a compound event may be associated with
more than one value of the random variable.
Statistics is concerned with making inferences about populations and population characteristics.
Experiments are conducted with results that are subject to chance. It is often important to allocate a
numerical description to the outcome. A random variable can be of the discrete or continuous type.
Discrete Random Variable
A discrete variable can only assume at most a countable set of isolated values, as already stated, such
as {0, 1, 2,3 4, 5}, {1, 2, 3,4, …}, { 1/2 , 1/3 , 5/7, 4/9 }, or { 0.1 , 0.15, 0.75, 1.45 } etc.
For example, the sample space giving a detailed description of each possible outcome when three
components are tested may be written
S = {NNN, NND, NDN, DNN, NDD, DND, DDN, DDD},
where N denotes non defective and D denotes defective. One is naturally concerned with the number
of defectives that occur. Thus, each point in the sample space will be assigned a numerical value of 0, 1, 2, or
3. These values are, of course, random quantities determined by the outcome of the experiment. They may be
viewed as values assumed by the random variable X, the number of defective items when three components
are tested.
we notice that the random variable X assumes the value 2 for all elements in the subset
E = {DDN, DND, NDD}
of the sample space S. That is, each possible value of X represents an event that is a subset of the sample
space for the given experiment.
(1) Probability Mass Function of a Discrete Random Variable
If a random variable X assumes discrete values (finite or countably infinite); x1, x2, … , xn, …… , it
is called discrete random variable. The function f(x) defined as
 P[X = xi] if x RX
PX(xi) =  where RX denotes the range of X
 0 if x RX
is called the probability function (or the probability mass function, pmf ) of the random variable X
with the properties:
(i) 0  PX(xi)  1

(ii)  PX(xi) = 1
i=1
(iii) PX(xi) = P(X = xi)
In the following examples, for defined random variables, we can find out the probability distribution.
Examples
(1) Two balls are drawn in succession without replacement from an urn containing 4 red balls and 3
black balls. Let Y denotes the number of red balls.
4th Semester MME, 2017

(2) Consider the simple condition in which components are arriving from the production line and they
are stipulated to be defective or not defective. Lets the random variable X by
0 if the component is not defective
X= 1
 if the component is defective
(3) The random variable for which 0 and 1 are chosen to describe the two possible values is called a
Bernoulli random variable.
(4) The probability distribution for the sample space for tossing three coins is
X: 0 1 2 3 Total
Probability P(X) 1/8 3/8 3/8 1/8 1
(5) The baseball World Series is played by the winner of the National League and the American League.
The first team to win four games wins the World Series. In other words, the series will consist of
four to seven games, depending on the individual victories. The data shown consist of the number of
games played in the World Series from 1965 through 2005. (There was no World Series in 1994.)
The number of games played is represented by the variable X. Find the probability P(X) for each X,
construct a probability distribution, and draw a graph for the data.
“Introductory Statistics”, by Bluman, Example 5.3
X: 4 5 6 7 Total
Number of Games Played 8 7 9 16 40
The probability P(X) can be computed for each X by dividing the number of games X by the total.
8 6
For 4 games, = 0.200 for 6 games = 0.225
40 40
7 16
For 5 games, 40 = 0.175 for 7 games 40 = 0.400
The probability distribution is

X: 4 5 6 7 Total
Number of Games Played 0.200 0.175 0.225 0.400 1
Examples (1)
Find the probability distribution of the sum of the dots when two dice are thrown.
Sol.
We are familiar with the sample space when two dice are thrown:
Let X denotes the sum of the dots of the top faces, so the range of
X = 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12
X = 2 means the event {(1,1)}, therefore P[X=2] =1/36
X = 3 means, the event {(1,2), (2,1)}, therefore P[X=3] = 2/36
X = 4 means the event {(1,3), (3,1), (2,2)}, therefore P[X=4] = 3/36
…. …. ….
X = 12 means the event {(6,6)}, therefore P[X=12] = 1/36
Hence the complete probability distribution is
Xi 2 3 4 5 6 7 8 9 10 11 12
PX(xi) = P[X=xi] 1/36 2/36 3/36 4/36 5/36 6/36 5/36 4/36 3/36 2/36 1/36

Exercise 1.
Many manufacturers have quality control programs that include inspection of incoming materials for
defects. Suppose a computer manufacturer receives computer boards in lots of five. Two boards are selected
from each lot for inspection. We can represent possible outcomes of the selection process by pairs. For
example, the pair (1, 2) represents the selection of boards 1 and 2 for inspection.
List the ten different possible outcomes.
Suppose that boards 1 and 2 are the only defective boards in a lot of five. Two boards are to be
chosen at random. Define X to be the number of defective boards observed among those inspected. Find the
probability distribution of X.
Exercise 2.
Let X be the number of years before a particular type of machines will need replacement. Assume
that X has the probability function f(1) = 0.1, f(2) = 0.2, f(3) = 0.2, f(4) = 0.2, f(5) = 0.3. Find the probability
that machine needs no replacement during the first 3 years.
(Kreyszig 7th Ex 24.5)
Exercise 3.
If X has the probability function f(x) = k/2x (X= 0,1,2,3,…) What is the value of k and find P(X  4)
(Kreyszig 7th Ex 24.5)
Cumulative Distribution Functions
It is sometimes useful to provide cumulative probabilities. Such probabilities can be used to find
the probability mass function of a random variable. Therefore, using cumulative probabilities is an alternate
method of describing the probability distribution of a random variable.
There are also many problems where we may wish to compute the probability that the observed
value of a random variable X will be less than or equal to some real number x. Writing F(x) = P(X ≤ x) for
every real number x, we define F(x) to be the cumulative distribution function of the random variable X.
Definition (Cumulative Distribution Function)
In the case of a discrete random variable, FX (x) is the sum of the probabilities of all possible values
of X that are less than or equal to the argument x. That is,
FX (x) =  PX(xk ) =  P[xk  x]
xk < x xk  x
This is summed over all possible xk less than or equal to x.
For a discrete random variable X, FX(x) satisfies the following properties
1. FX(x) = P(X  x)
2. 0  FX(x)  1
Examples (2)
The number of floods recorded per year at a gauging station in Italy is given in the following table.
For this data, the pmf is as follows: (A flood occurrence is defined as river discharge exceeding 300 m3 / s)
pX (0) = .00; pX (1) = 2/34 = .06; pX (2) = 6/34 = .18; pX (3) = 7/34 = .20; pX (4) = 9/34 = .26;
pX (5) = 4/34 = .12; pX (6) = 1/34 = .03; pX (7) = 4/34 = .12; pX (8) = 1/34 = .03; pX (x) = .00 for x > 8

Number of Floods Number of A plot of probability mass function.

in a year occurrences
0 0
1 2
2 6
3 7
4 9
5 4
6 1
7 4 In some text, this also called probability histogram
8 1
9 0
Total = 34
.
Cumulative distribution function (cdf) of flood occurrences is:
FX (0) = 0.00; FX (1) = 0.06; FX (2) = 0.24; FX (3) = 0.44; FX (4) = 0.70;
FX (5) = 0.82; FX (6) = 0.85; FX (7) = 0.97; FX (8) = 1.0; FX (x) = 1.00, for x > 8.
Examples (3)
Suppose that a day‟s production of 850 manufactured parts contains 50 parts that do not conform to
customer requirements. Two parts are selected at random, without replacement, from the batch. Let the
random variable X equal the number of nonconforming parts in the sample. What is the cumulative
distribution function of X? (Example 3.8, “Probability and Statistics for Engineers” by Douglas and Montgomery)
Solution
First find the probability mass function of X.
800 799
P(X = 0) = 850  849 = 0.886,
800 50
P(X = 1) = 2  850 849 = 0.111,
50 49
P(X = 2) =  = 0.003
850 849
Using probability mass function, we find the cumulative distribution function at each point. Therefore,
F(0) = P(X  0) = 0.886,
F(1) = P(X  1) = 0.997,
F(2) = P(X  2) = 1
Examples (4)
If a car agency sells 50% of its inventory of a certain foreign car equipped with side airbags, find a
formula for the probability distribution of the number of cars with side airbags among the next 4 cars sold by
the agency. (Walpole 9th Example 3.9)

Solution
Since the probability of selling an automobile with side airbags is 0.5, the 24 = 16 points in the
sample space are equally likely to occur. Therefore, the denominator for all probabilities, and also for our
function, is 16. To obtain the number of ways of selling 3 cars with side airbags, we need to consider the
number of ways of partitioning 4 outcomes into two cells, with 3 cars with side airbags assigned to one cell
and the model without side airbags assigned to the other. This can be done in 4C3 = 4 ways. In general, the
event of selling x models with side airbags and 4−x models without side airbags can occur in 4Cx ways,
where x can be 0, 1, 2, 3, or 4. Thus, the probability distribution f(x) = P(X = x) is
1
f(x) = 16 4Cx , for x = 0, 1, 2, 3, 4.
Examples (5)
Complete probability distribution is:
X: 0 1 2 3 4
f(x) = P[X = x] 1/16 1/4 3/8 1/4 1/16
Now Using F(x), verify that f(2) = 3/8.
F(0) = f(0) = 1/16
F(1) = f(0) + f(1)= 5/16
F(2) = f(0) + f(1) + f(2) = 11/16
F(3) = f(0) + f(1) + f(2) +f(3) = 15/16
F(4) = f(0) + f(1) + f(2) + f(3) + f(4) = 1
Now f(2) = F(2) – F(1) = 11/16 – 5/16 = 3/8
Exercise 4.
Errors in an experimental transmission channel are found when the transmission is checked by a
certifier that detects missing pulses. The number of errors found in an eight bit byte is a random variable with
the following distribution: (Montgomery Douglas 5th Ed. Problem 3-40)

0 if x < 1
if 1  x < 4

0.7
PX(x) =
0.9 if 4  x < 7
 1 if 7  x
Determine each of the following probabilities:
(a) P(X  4), (b) P(X > 7), (c) P(X  5), (d) P(X > 4), (e) P(X  2)
(2) Mean and Variance of a Discrete Random Variable
Two numbers are often used to summarize a probability distribution for a random variable X. The
mean is a measure of the center or middle of the probability distribution, and the variance is a measure of the
dispersion, or variability in the distribution.
Definition
The mean or expected value of a random variable X indicates its weighted mean of all of its possible
values x1,x2, … ,xn each weighted by the respective probability.
Let X be a discrete random variable having values x1, x2, … , xn with corresponding probabilities
P(X=x1) , P(X=x2), … , P(X=xn) such that  P(X=xi) = 1. Then the mathematical expectation or the
expectation or the expected value of X, denoted by E(X) is defined as

n
E(X) = x1 P(X=x1) + x2 P(X=x2) + … + P(X=xn) xn =  xi P(X=xi)
i =1
The measure of variability of a random variable X is referred to as the variance of the random
variable or the variance of the probability distribution of X is denoted by Var (X) and defined as:
n
Var(X) = E[(X  )2 ] =  (X  )2 P(X=xi) , if X is discrete and
i =1
Alternatively
Let X be a continuous random variable with E(X)=. Then
n
2 = E[(X-)2] =  (x-)2 P(X=xi)
i =1
n
=  (x2-2x+2) P(X=xi)
i =1
n n n
=  x2 P(X=xi) -2  x P(X=xi) +2  P(X=xi)
i =1 i =1 i =1
2 2
= E(X )-2 E(X)+ .1
= E(X2)-22+2
= E(X2)-2
= E(X2)  [E(X)]2
Examples (6)
Two new product designs are to be compared on the basis of revenue potential. Marketing feels that
the revenue from design A can be predicted quite accurately to be $3 million. The revenue potential of design
B is more difficult to assess. Marketing concludes that there is a probability of 0.3 that the revenue from
design B will be $7 million, but there is a 0.7 probability that the revenue will be only $2 million. Which
design do you prefer?
Solution
Let X denote the revenue from design A. Because there is no uncertainty in the revenue from design
A, we can model the distribution of the random variable X as $3 million with probability 1. Therefore, E(X)
= $3 million.
Let Y denote the revenue from design B. The expected value of Y in millions of dollars is
E(Y) = $ 7(0.3) + $2(0.7) = $3.5
Because E(Y) exceeds E(X), we might prefer design B. However, the variability of the result from
design B is larger. That is,
2 = (7 – 3.5)2 (0.3) + (2 – 3.5)2 (0.7)
= 5.25 millions of dollars squared
 Standard Deviation =  = 2 = $2.29
Because the units of standard deviation are the same as the units of the random variable, the standard
deviation is easier to interpret. In this example, we can summarize our results as “the average deviation of Y
from its mean is $2.29 million.‟
Examples (7)
How many heads would you expect if you flipped a coin twice?
X = number of heads = {0, 1, 2} with p(0) = ¼, p(1) = ½ and p(2) = ¼

Weighted Average = E(X) = 0¼ + 1½ + 2¼ = 1

and E(X2) = 02¼ + 12½ + 22¼ = 3/2
Then Var (X) = E(X2) [E(X)]2 = 3/2 1 = ½
Examples (8)
Roll a die. If the side comes up is odd, you win the $ equivalent of that side. If it is even, you lose $ 4.
Let X denotes our earning, so the possible values of the variable X are: 1,3,5, -4, therefore
X 1 3 5 -4
P(X=x) 1/6 1/6 1/6 1/6
E(X) = 11/6 + 31/6 + 51/6 + (-4)  3/6 = - ½
E(X2) = 121/6 + 321/6 + 521/6 + (-4)2  3/6 = 1/6 + 9/6 + 25/6 + 48 / 6 = 13.8
Var (X) = 13.8  0.25 = 13.55
Exercise 5.
If it rains, an umbrella salesman can earn $30 per day. If it is fair he can lose $6 per day. What is his
expectation, if the probability of rain is 0.3?
Exercise 6.
A coin is biased such that a head is three times as likely to occur as a tail. Find the expected number
of tails when this coin is tossed twice. (Walpole Problem 4.4)
Exercise 7.
A man draws two balls from a bag containing 3white and 5 black balls. If he receives Rs.70 for every
white ball he draws and Rs.7 for every black ball, find his expectation? 2014
Exercise 8.
In summer season, a dealer of desert room cooler can earn Rs.500 per day if the day is hot and can
earn Rs.200 per day if it is fair and loses Rs.60 per day if it is cloudy. Find the expectation if the probability
of the day being hot is 0.40, for being fair it is 0.35 and being cloudy it is 0.25?
Exercise 9.
One thousand tickets are sold at $1 each for a color television valued at $350. What is the expected
value of the gain if you purchase one ticket?
Example 5.12, page 264, “Introductory Statistics” by Bluman
Exercise 10.
One thousand tickets are sold at $1 each for four prizes of $100, $50, $25, and $10. After each prize
drawing, the winning ticket is then returned to the pool of tickets. What is the expected value if you purchase
two tickets?
Example 5.13, page 265, “Introductory Statistics” by Bluman
Examples (9)
Two new product designs are to be compared on the basis of revenue potential. Marketing feels that
the revenue from design A can be predicted quite accurately to be $3 million. The revenue potential of design
B is more difficult to assess. Marketing concludes that there is a probability of 0.3 that the revenue from
design B will be $7 million, but there is a 0.7 probability that the revenue will be only $2 million. Which
design do you prefer? (Douglas Montgomery Example 3.10)

Solution
Let X denote the revenue from design A. Because there is no uncertainty in the revenue from design
A, we can model the distribution of the random variable X as $3 million with probability 1. Therefore,
E(X) = $3 million.
Let Y denote the revenue from design B. The expected value of Y in millions of dollars is
E(Y ) = $7(0.3) + $2(0.7) = $3.5
Because E(Y) exceeds E(X), we might prefer design B. However, the variability of the result from
design B is larger. That is,
2 = (7  3.5)2 (0.3) + (2  3.5)2 (0.7) = 5.25 millions of dollars squared
Because the units of the variables in this example are millions of dollars, and because the variance of
a random variable squares the deviations from the mean, the units of are millions of dollars squared. These
units make interpretation difficult. Because the units of standard deviation are the same as the units of the
random variable, the standard deviation is easier to interpret. In this example, we can summarize our results
as “the average deviation of Y from its mean is $2.29 million.‟‟
(3) Discrete Uniform Distribution
A random variable X has a discrete uniform distribution if each of the n values in its range, say,
x1, x2, … , xn with equal probability. Then
f(x) = 1/n
Suppose the range of the discrete random variable X is the consecutive integers a, a+1, a+2, …, b,
for a < b. The range of X contains b – a + 1 values each with probability 1 / (b-a+1).
Now
b 1
 =  k (b-a+1 )
k =a
b b(b+1) - (a-1)a
The algebraic identity  k = 2 can be used to simplify the result to
k =a
b+a
= 2
and the variance of X is
(b-a+1)2 - 1
2 = 12 (verify!)
Examples (10)
The first digit of a part‟s serial number is equally likely to be any one of the digits 0 through 9. If one
part is selected from a large batch and X is the first digit of the serial number, X has a discrete uniform
distribution with probability 0.1 for each value in . That is, for each value in R. The probability mass function
of X is shown in Figure.
Which is a probability mass function for a discrete uniform random variable

Examples (11)
A voice communication system for a business contains 48 external lines. At a particular time, the
system is observed, and some of the lines are being used. Let the random variable X denote the number of
lines in use. Then, X can assume any of the integer values 0 through 48. Then,
E(X) = (48 + 0)/2 = 24
And  = {[(48 – 0 + 1)2  1]/12}1./2 = 14.14
Practical Interpretation:
The average number of lines in use is 24 but the dispersion (as measured by ) is large. Therefore,
at many times far more or fewer than 24 lines are in use.
Exercise 11.
Suppose that X has a discrete uniform distribution on the integers 0 through 9. Determine the mean, variance,
and standard deviation of the random variable Y = 5X and compare to the corresponding results for X.
(Problem 3.71, page 79 “Probability and Statistics for Engineers” by Douglas and Montgomery)
Exercise 12.
Thickness measurements of a coating process are made to the nearest hundredth of a millimeter. The
thickness measurements are uniformly distributed with values 0.15, 0.16, 0.17, 0.18, and 0.19. Determine the
mean and variance of the coating thickness for this process.
(Problem 3.49, page 71 “Probability and Statistics for Engineers” by Douglas and Montgomery)
Now we discuss some well known probability distributions
(1) Bernoulli Distribution

(i). Bernoulli Experiment:
Suppose a random experiment has two outcomes, namely Success and Failure. Let
probability of success be p and probability of failure be q = 1  p. Such an experiment is called
Bernoulli experiment or Bernoulli trial.
Trials of random experiment are called Bernoulli trials, if they satisfy the following conditions:
1. There is finite number of trials
2. The trials should be independent.
3. Each trial has exactly two outcomes: success or failure.
4. The probability of success remains the same in each trial.
(ii). Bernoulli Distribution:
Bernoulli distribution has one constant, namely p. This constant p is called a parameter of
the Bernoulli distribution.
Bernoulli distribution can also be written as P(x) = px q1−x, x = 0,1
Here, occurrence of the value 1 may be termed as success and the occurrence of the value 0
may be termed as failure.
Therefore P (success) = p and P(failure) = q = 1  p.
We denote the Bernoulli random variable X ~ BER(p) where p is its parameter.
We can calculate its expectation and variance as follows.
E(X) = 0  (1  p) + 1  p = p
E(X2) = 02  (1  p) + 12  p = p
Var (X) = E(X2)  [E(X)]2
= p  p2
= p(1  p)
= pq
Examples
1. A fair coin is tossed. Let the variable x take values 1 and 0 according to as the toss results in
„Head„ or „Tail‟. Then, X is a Bernoulli variable with parameter p = 1/2. Here, X denotes the
number of heads obtained in the toss. Probability of success = 1/2 and the probability of
failure = 1/2.
2. In a single throw of a dice, the outcome "5" is called a success and any other outcome is
called a failure, then the successive throws of a dice will contain Bernoulli trials. The
probability of success = 1/6 and the probability of failure = 5/6.
Examples (12)
What is the probability of getting a score of not less than 5 in a throw of a six-sided die?
Solution
Although there are six possible scores {1, 2, 3, 4, 5, 6}, we are grouping them into two sets, namely
{1, 2, 3, 4} and {5, 6}. Any score in {1, 2, 3, 4} is a failure and any score in {56} is a success. Thus, this
is a Bernoulli trial with
P(X = 0) = P(failure) = 4/6,
and P(X = 1) = P(success) = 2/6
Hence, the probability of getting a score of not less than 5 in a throw of a six-sided die is 2/6.
(2) Binomial Probability Distribution
Consider a fixed number of mutually independent Bernoulli trails. Suppose these trials have same
probability of success, say p. A random variable X is called a binomial random variable if it represents the
total number of successes in n independent Bernoulli trials.
Thus, to find the probability density function of X we have to find the probability of successes in n
independent trails. If we have successes in trails, then the probability of each n-tuple with x successes
and n − x failures is
px (1 − p)n−x.
However, there are  x  tuples with x successes and n − x failures in n trials.

n
Hence P[X = x] =  x  p q
n x nx
( x = 0, 1, 2, … ,n )
Therefore the probability density function of X is
f(x) =  x  p q
n x nx
( x = 0, 1, 2, … ,n )
The random variable X is called the binomial random variable with parameters and if its
probability density function is of the form
f(x) =  x  p q
n x nx
( x = 0, 1, 2, … ,n )

Remarks
The Bernoulli distribution is closely related to the Binomial distribution. As long as each individual
Bernoulli trial is independent, then the number of successes in a series of Bernoulli trails has a Binomial
Distribution. The Bernoulli distribution can also be defined as the Binomial distribution with n=1.
Examples (13)
On a five-question multiple-choice test there are five possible answers, of which one is correct. If a
student guesses randomly and independently.
(i) what is the probability that she is correct only on questions 1 and 4?
(ii) what is the probability that she is correct only on two questions?
Solution (i)
1 4
The probability of a success is p = , and thus 1  p = , therefore, the probability that she is
5 5
correct on question 1 and 4 is
2 3
P(correct on question 1 and 4) = p (1  p) =  5   5  = 55 = 0.02048
2 1 3 4 64
   
Solution (ii)
2 3
P(correct on any two questions) = C2 p (1  p) = 10   5   5  = 5 = 0.2048
5 2 31 4 64
    5
Mean and Variance
Theorem
The mean and variance of the binomial distribution b(x; n, p) are
 = np and 2 = npq
Proof
Let X be a random variable with the binomial distribution b(x; n, p). Then its mean and variance are
given by  = np and 2 = npq respectively.
Now Mean,  = E(X)
=  x  x  pxqn-x,
n n
where x = 0,1,2, … ,n.
x=0
= 0.qn + 1.  1  p1qn-1 + 2.  2  p2qn-2 + …+ npn

n n
= np [qn-1 +  1  pqn-2 +  2  p2qn-3 + …+ pn-1

n-1 n-1
= np (q + p)n-1 = np , because q + p = 1
Hence mean = np. In other words, the mean number of success is np. Similarly the mean number of
failures is nq.
By definition the variance 2, is given by
2= E[ X - ]2 = E(X2) – [E(X)]2
but E(X2) = E[X (X-1) + X]
= E[X (X - 1)] + E(X)
= E[X (X - 1)] + np

E[X (X - 1)] =  x (x – 1)  x  pxqn-x

n n
Now
x=0
n n (n - 1) (n - 2)!
= 0 + 0 +  x (x - 1) x (x - 1) (x - 2)! (n - x)! p2 px–2qn-x
x=2
n (n - 2)!
= n (n - 1) p2  (x - 2)! (n - x)! px–2qn-x
x=2
The term (n – x) may be written as [(n – 2) – (x - 2)]

Now, by substituting y = x – 2 and m = n – 2 in the summation, we obtain
m m!
E[X (X - 1)] = n (n - 1) p2  y! (m - y)! pyqm-y
y=0
= n (n - 1) p2   y  pyqm-y
m m
y=0
= n (n - 1) p2 (because summation is 1)
Thus  = E(X ) – [E(X)] = E[X (X - 1)] + E(X) - [E(X)]2

2 2 2
= n (n - 1) p2 + np – (np)2
= n2p2 – np2 + np – n2p2 = np – np2
= np(1 - p) = npq (because q + p = 1; i.e. 1 – p = q)
and  = npq
Hence the variance of the number of success is npq, and the standard deviation is npq .
Examples (14)
Tests for impurities commonly found in drinking water from private wells showed that 30% of all
wells in a particular county have impurity A. If a random sample of five wells is selected from the large
number of wells in the county, what is the probability that:
a. Exactly three will have impurity A?
b. At least three?
c. Fewer than three?
Solution
The sampling procedure represents a Binomial experiment with n = 5 and p = 0.3.
a. The probability of drawing exactly y = 3 wells containing impurity A is
p(y) =  y  p q
n y ny
when n = 5, p = 0.3 and y = 3, Thus

5!
p(3) = 3!2! (0.3)3 (0.7)2 = 0.13230
b. The probability of observing at least three wells containing impurity A is

P(y  3) = p(3) + p(4) + p(5)
5!
Where p(4) = 4!1! (0.3)4 (0.7)1 = 0.02835

5!
p(4) = 5!0! (0.3)5 (0.7)0 = 0.00243
Since we found p(3) in part a. we have

P(y  3) = p(3) + p(4) + p(5) =0.13230 + 0.02835 + 0.00243
c. P(y < 3) = p(0) + p(1) + p(2)
by using the complementary relationship and the fact that  P(y) = 1, we have
p(y < 3) = 1  p(y  3) = 1  0.16308 = 0.83692
Examples (15)
A large chain retailer purchases a certain kind of electronic device from a manufacturer. The
manufacturer indicates that the defective rate of the device is 3%.
(a) The inspector of the retailer randomly picks 20 items from a shipment. What is
the probability that there will be at least one defective item among these 20?
(b) Suppose that the retailer receives 10 shipments in a month and the inspector randomly tests
20 devices per shipment. What is the probability that there will be 3 shipments containing at
least one defective device?
(Example 5.6, Book: Probability and Statistics by Walpole 8th Ed)
Solution (a)
Denote by X the number of defective devices among the 20. This follows that
n = 20 , X = 0, 1, 2, …, 20 and p = 0.03,
the probability of a defective device, therefore
P(X > 1) = 1 - P(X = 0) = 1 – 20 C0 p0 q20-0
= 1 - 0.03°(1 - 0.03)20 = 0.4562.
Solution (b)
In this case, each shipment contains at least one defective. Hence, testing the result of each shipment
is a trial with p = 0.4562 from part (a). Assuming the independence from shipment to shipment and denoting
by Y the number of shipments containing at least one defective item Y follows another binomial distribution
b(x; 10, 0.4562). Therefore, the answer to this question is
10 3 7
P(Y = 3) = C3 0.4562 (1 - 0.4562)
= 0.1602
Moment Generating Function (mgf)
Let X be a random variable with probability function (x) and a  x  b or -  x  +.
-
Let t be a real valued parameter which is independent of x. We define e tx , et(x-c), et(x-x) to be the moment
generating function about the origin, about a fixed point c and about the mean of the distribution x-
respectively. These are denoted by M0(t), Mc(t) and Mx- (t) respectively.
M0(t) = E(etx)
Mc(t) = E(et(x-c))
-
Mx- (t) = E(et(x-x))
dn dn dn -
and dt n M0(t),
dt n Mc(t) and
dtn Mx(t) are the nth moment about origin, about c and about the
mean of the distribution respectively.

Mean and Variance Using mgf

Find the moment generating function M0(t) of a binomial probability distribution. Also find the mean
and variance using this M0(t).
Sol. The moment generating function M0(t) about the origin is
M0(t) = E(etx)
n n
=  etxf(x) or =  etx b(x; n, p)
x=0 x=0
n
=  etx n
Cx qn-x px
x=0
n
=  n
Cx(et p)x qn-x
x=0
n
=  n
Cx qn-x (et p)x
x=0
= (q + et p)n
i.e. M0(t) = (q + e p)n
t
………………… (i)
Diff. (i) w r t „t‟ and putting t = 0
d d
1 = E(X) = dt M0(t) = dt (q + et p)n
= n (q + et p)n-1 et p ……… (ii)
= n (q + p)n-1 p putting t = 0
n-1
= np (1)
=np which is the mean value.
again diff (ii) w r t „t‟ and putting t = 0
d
2 = E(X2) = dt { n (q + et p)n-1 et p }
= n (q + et p)n-1 et p + n(n-1) (q + et p)n-2 (et p)2

= n (q + p)n-1 p + n(n-1) (q + p)n-2 p2 putting t = 0
= np + n(n-1) p = n p + n p  n p
2 2 2 2
Now Var(X) = 2 = 2 - 21 = n p + n2 p2  n p2  (np)2

= n p  n p2
= n p (1  p) = n p q which is the variance.
(3) The Poisson Probability Distribution
A discussion
Flaws occur at random along the length of a thin copper wire. Let X denote the random variable that
counts the number of flaws in a length of L millimeters of wire and suppose that the average number of flaws
in L millimeters is . Partition the length of wire into n subintervals of small length, say, 1 micrometer each.
If the subinterval chosen is small enough, the probability that more than one flaw occurs in the subinterval is
negligible.
Furthermore, we can interpret the assumption that flaws occur at random to imply that every
subinterval has the same probability of containing a flaw, say, p. Finally, if we assume that the probability
that a subinterval contains a flaw is independent of other subintervals, we can model the distribution of X as
approximately a binomial random variable. Because

E(X) =  = np, we obtain p = /n. That is, the probability that a subinterval contains a flaw is /n.
with small enough subintervals, n is very large and p is very small.
The same reasoning can be applied to any interval, including an interval of time, an area, or a volume.
For example, counts of (1) particles of contamination in semiconductor manufacturing, (2) flaws in rolls of
textiles, (3) calls to a telephone exchange, (4) power outages, and (5) atomic particles emitted from a
specimen have all been successfully modeled by the probability mass function in the following definition
Definition
Given an interval of real numbers, assume counts occur at random throughout the interval. If the
interval can be partitioned into subintervals of small enough length such that
(1) the probability of more than one count in a subinterval is zero,
(2) the probability of one count in a subinterval is the same for all subintervals and
proportional to the length of the subinterval, and
(3) the count in each subinterval is independent of other subintervals,
the random experiment is called a Poisson process.
The random variable X that equals the number of counts in the interval is a Poisson random variable with
parameter 0 < , and the probability mass function of X is
xe-
P(X ; ) = x! , x = 0, 1, 2, …
The characteristics of a Poisson random variable are listed below:

the experiment consist of counting the number X of times a particular event occur during a
given unit of time, or in a given area of volume (or weight, distance, or any other unit of
measurement).
The probability that an event occurs in a given unit of time, area, or volume is the same for
all the units.
The number of events that occur in one unit of time, area, or volume is independent of the
number that occurs in other units.
The mean (or expected) number of events in each unit will be denoted by the Greek letter
meu, 
The Mean and Variance
Mean =  and 2 =  (Verify !)
 is called the parameter of a Poisson distribution.
Poisson Approximation to the Binomial
To derive an approximate formula to the binomial distribution b(x; n, p) when n, p0, and the
product np remains constant, we proceed as below:
The binomial distribution b(x; n, p) may be written as
b(x; n, p) =  x  pxqn-x
n
for n = 0,1,2,3, …
n(n - 1)(n - 2) … (n - x + 1) x n-x
= x! pq
 
Let np = . Then p = n and q = 1 – p = 1 - n
Replacing all terms involving p, we get

x n-x
n(n - 1)(n - 2) … (n - x + 1)     
b(x; n, p) = 1 n
x! n  
n x
x n(n - 1)(n - 2) … (n - x + 1)   
= 1   1  
x! nx  n  n
n x
x  
= x! [1. 1  n  1  n  … 1  n ] 1  n  1  n 
1 2 x1
        
let n   and p  0 so that np =  remains constant, we observe that each of the term 1  n  ,
1
 
x
1  2  , … , 1  x1  and 1    approaches unity.
 n  n   n
n n/
 
The term 1  n  may be written as [1  n  ]
   
k
 
Writing n/ = k, the expression becomes [1  n  ].
 
k

If n increases indefinitely, so does k. Therefore 1  n  tends to e1, where e = 2.71828,
 
n

therefore, lim 1   = e
n  n
Thus the limiting value of p(X = x) is given by the expression
x xe-
lim b(x; n , p) = x! . 1 . 1 … 1 . e = x! , for x = 0, 1, … , 
n
In other words, if X is a Binomial random variable such that

n
p (X = x) =  x  pxqn-x for x = 0,1,2,3, …
xe-
lim p(X = x) = x! , x = 0, 1, 2, … , 
n
np
It is denoted by p(x; ). Hence a random variable X having probability distribution p(x; ) is said to
have a Poisson distribution with parameter .
Examples (16)
In a certain industrial facility, accidents occur infrequently. It is known that the probability of an
accident on any given day is 0.005 and accidents are independent of each other.
(a) What is the probability that in any given period of 400 days there will be an accident on one day?
(b) What is the probability that there are at most three days with an accident? (Walpole Example 5.19)
Solution
Let X be a binomial random variable with n = 400 and p = 0.005. Thus, np = 2.
Using the Poisson approximation,
(a) P(X = 1) = e−221 = 0.271 and
3
(b) P(X ≤ 3) =  e-22x/x! = 0.857
x=0

Examples (17)
In a manufacturing process where glass products are made, defects or bubbles occur, occasionally
rendering the piece undesirable for marketing. It is known that, on average, 1 in every 1000 of these items
produced has one or more bubbles. What is the probability that a random sample of 8000 will yield fewer
than 7 items possessing bubbles?
Solution
This is essentially a binomial experiment with n = 8000 and p = 0.001. Since p is very close to 0 and
n is quite large, we shall approximate with the Poisson distribution using
 = (8000)(0.001) = 8.
Hence, if X represents the number of bubbles, we have
6
P(X < 7) =  b(x; 8000, 0.001) ≈ p(x;8) = 0.3134.
x=0
Exercise 13.
Suppose that, on average, 1 person in 1000 makes a numerical error in preparing his or her income
tax return. If 10,000 returns are selected at random and examined, find the probability that 6, 7, or 8 of them
contain an error. (Walpole 5.61) Ans. 0.2657
Exercise 14.
Potholes on a highway can be a serious problem, and are in constant need of repair. With a particular
type of terrain and make of concrete, past experience suggests that there are, on the average, 2 potholes per
mile after a certain amount of usage. It is assumed that the Poisson process applies to the random variable “
number of potholes.” (Walpole Problem 5.72)
(a) What is the probability that no more than one pothole will appear in a section of 1 mile?
(b) What is the probability that no more than 4 potholes will occur in a given section of 5 miles?
Mean and Variance Using mgf (optional)
Find the moment generating function M0(t) of a Poisson probability distribution. Also find the mean
and variance using this M0(t).
Sol. The moment generating function M0(t) about the origin is
n n
M0(t) = E(etx) =  etxf(x) or =  etx b(x; n, p)
x=0 x=0
Examples (18)
If there are 200 typographical errors randomly distributed in a 500-page manuscript, find the
probability that a given page contains exactly 3 errors. (Bluman Example 5.27)
Solution
First, find the mean number  of errors. Since there are 200 errors distributed over 500 pages, each
page has an average of
200 2
 = 500 = 5 = 0.4
xe- (0.4)3e-0.4
P(X; ) = = = 0.0072
x! 3!
Thus, there is less than a 1% chance that any given page will contain exactly 3 errors.

Examples
For the case of the thin copper wire, suppose that the number of flaws follows a Poisson distribution
with a mean of 2.3 flaws per millimeter.
(i) Determine the probability of exactly two flaws in 1 millimeter of wire.
Let X denote the number of flaws in 1 millimeter of wire. Then, E(X) = 2.3 flaws and
e-2.3 (2.3)2
P(X = 2) = 2! = 0.265
(ii) Determine the probability of 10 flaws in 5 millimeters of wire.
Let X denote the number of flaws in 5 millimeters of wire. Then, X has a Poisson distribution with
E(X) = 5 mm  2.3 flaws/mm = 11.5 flaws
Therefore
e-11.5 (11.5)10
P(X = 10) = 10! = 0.113
(iii) Determine the probability of at least one flaw in 2 millimeters of wire.
Let X denote the number of flaws in 2 millimeters of wire. Then, X has a Poisson distribution with
E(X) = 2 mm  2.3 flaws/mm = 4.6 flaws
Therefore,
P(X  1) = 1  P(X = 0) = 1  e-4.6 = 0.9899
Practical Interpretation: Notice that when a probability was requested for 2 mm of wire 2 was
adjusted to 4.6, the mean number of flaws in 2 mm. With such adjustments, probabilities can be calculated
for intervals of any size.
Examples (19)
Suppose that number Y of a company‟s employees absent on Mondays has approximately a Poisson
probability distribution. Furthermore, assume that the average number of Monday absent is 2.5.
Find the mean and the standard deviation of Y, the number of employees absent on Monday.
Find the probability that exactly 5 employees are absent on a given Monday.
Find the probability that 2 or more employees are absent on a Monday.
Solution
a) The mean and variance of a Poisson random variable are both equals to . Thus, for this example
 =  = 2.5 and 2 =  = 2.5
then the standard deviation is
 = 2.5 = 1.58
b) We want the probability that exactly 5 employees are absent on Monday.
ye-
The probability distribution for y is P(y) = y!
Then, since  = 2.5, y = 5 and e-2.5 = 0.082085
(2.5)5e-2.5 (2.5)2 (0.082085)
 P(5; 2.5) = 5! = = 0.067
5.4.3.2.1
c) to find the probability that two or more employees or absent on Monday,
we need to find

P(y 2) = P(2) + P(3) + P(4) + … =  P(y)
y=2
In order to find the probability of this event, we must consider the complementary event.

Thus, P(y 2) = 1 – P(y  1) = 1 – [P(0) + P(1)]

(2.5)0e-2.5 (2.5)1e-2.5 1 . (0.082085) (2.5)2 (0.082085)
=1- 0! - 1! = 1 - 1 - 1
= 1 – 0.287
= 0.713
According to our Poisson model, the probability that two or more employees are absent on Monday is 0.713.
Examples
(i) If on the average 2 cars enter a certain parking lot per minute, what is the probability that during any
given minute 4 or more cars will enter the lot? Ans. 14.3% (Kreyszig 10th Ed)
(ii) A business receives order at an average rate of 1 per minute. What is the probability of getting three
orders in one minute? (0.0613 or 6%)
(iii) An emergency service receives an average of 2.1 false alarms per day. What is the probability of
getting four false alarms in a given day? (0.0992 or 10%)
(iv) On average the demand for a certain product is four per week. If the stock at the beginning of each
week is renewed so that there are always 6 in store, what is the probability of running out of stock in
any week (13.4 %)
(v) A call centre has a capacity to deal with 25 calls per minute on average. What is the probability of
getting 30 calls in any minute period? (4.5%)
(vi) The average time taken for a worker to assemble a certain product is 45 minutes. There are 10
workers employed to make these assemblies. What is the probability of assembling 10 units in an
hour? (8%)
Examples (20)
The probability that a man aged 50 years will die within a year is 0.01125. What is the probability
that of 12 such men at least 11 will reach their fifty-first birthday?
Solution
Here p = 0.01125 and n = 12. We compute the desired probability by means of Poisson distribution
because the probability of death is very small.
Therefore  = np = 12  (0.01125) = 0.135, and the Poisson distribution is
e-0.135(0.135)x
P(x; 0.135) = x!
Now the probability that no person will die, i.e. all the 12 persons will survive, is
(0.135)2 (0.135)3
P(0; 0.135) = e-0.135 = 1  0.135 + 2!  3! + … = 0.8737,
and the probability that 1 person will die, i.e. 11 persons will survive, is
e-0.135(0.135)1
p(1; 0.135) = 1!
= (0.8737) (0.135) = 0.1179
Hence the probability that at least 11 persons will survive
= P(0; 0.135) + p(1; 0.135) = 0.8737 + 0.1179 = 0.9916
Examples (21)
Show that the expected value of a Poisson random variable is .
Solution
By definition, we have

 ye-
P(y) =  y P(y) =  y  y!
all y y=0
The first term of this series will equal 0, because y = 0. Therefore,
 y ye-  ye-   y-1e-
E(y) =  y! =  = 
y=1 y = 1 (y - 1)! y = 1 (y - 1)!
Factoring the constant y out side the summation and letting z = (y – 1), we obtain
 ze- 
E(y) =   z! =   P(z)
z=0 z=0
Where z is a Poisson random variable with mean . Hence,

E(y) =   P(z) =  (1) = 
z=0
Exercise 15.
Ten is the average number of oil tankers arriving each day at a certain port. The facilities at the port
can handle at most 15 tankers per day. What is the probability that on a given day tankers have to be turned
away? (Walpole 9th Example 5.18) And 0.0487
Exercise 16.
An automobile manufacturer is concerned about a fault in the braking mechanism of a particular
model. The fault can, on rare occasions, cause a catastrophe at high speed. The distribution of the number of
cars per year that will experience the catastrophe is a Poisson random variable with = 5.
(a) What is the probability that at most 3 cars per year will experience a catastrophe?
(b) What is the probability that more than 1 car per year will experience a catastrophe?
(Walpole 9th , Problem 5.65) Ans. (a) 0.2650; (b) 0.9596

CONTINUOUS RANDOM VARIABLES AND THEIR PROBABILITY DISTRIBUTIONS

(1) Continuous Random Variables
Random variables that can assume any one of the uncountable infinite numbers of points in one or
more intervals on the real line are called continuous random variables. The daily rainfalls at some location,
the strength (in pounds per square inch) of a steel bar, and the intensity of sunlight at a particular time of the
day are the examples of continuous random variables.
If a random variable X assumes every possible value in an interval [a, b] or (-, +) is called
continuous random variable. The function f(x) is called the probability density function or simply the density
function of the random variable X with the properties:
(i) 0  f(x)  1
+
(ii)  f(x) = 1

Expected Value of a Continuous Random Variable

Suppose X is a continuous random variable with probability density function f(x). The mean
or expected value of X, denoted as or E(X), is

E(X)=  x f(x)dx

The measure of variability of a random variable X is referred to as the variance of the random
variable or the variance of the probability distribution of X is denoted by Var (X) and defined as:

Var(X) = E[(X  )2 ] =  (X  )2 f (x) = E(X2)  [E(X)]2

Exercise 17.
The thickness of a conductive coating in micrometers has a density function of 600x2
for 100 m  x  120 m.
(a) Determine the mean and variance of the coating thickness.
(b) If the coating costs $0.50 per micrometer of thickness on each part,
what is the average cost of the coating per part? (Douglas Montgomery, Prob. 4.27)
(2) Continuous Uniform Distribution
The simplest continuous distribution is Continuous Uniform Distribution
A continuous random variable X with probability density function
1
f(x) = b - a , a  x  b
is a continuous uniform random variable.
The probability density function of a continuous uniform
random variable is shown in Figure
Mean and Variance:

The mean of the continuous uniform random variable X is
b
x 0.5x2 b (b - a)
E(X) =  b - a dx = b - a | = 2
a a

The variance of X is;

a+b a+b
b (x - ( 2 ))2 (x - ( 2 ))3 b
(b - a)2
Var (X) =  dx = |=
a b-a 3(b - a) a 12
Examples (22)
Let the continuous random variable X denote the current measured in a thin copper wire in
milliamperes. Assume that the range of X is [0, 20 mA], and assume that the probability density
function of X is f(x) =0.05, for 0  x  20. What is the probability that a current measurement is less
than 10 milliamperes?
(Example 4.1, “Probability and Statistics for Engineers” by Douglas and Montgomery)
Solution
The probability density function is shown in Figure. It is assumed that wherever it is not specifically
defined. The probability requested is indicated by the shaded
area in Figure.
10 10
P(X < 10) =  f(x) dx =  0.05 dx = 0.5

0 0
As another example,
20 20
P[5 < X < 20] =  f(x) dx =  0.05 dx = 0.5

5 5
Exercise 18.
The probability density function of the length of a metal rod is f(x) = 2, for 2.3 < x < 2.8 meters. If
the specifications for this process are from 2.25 to 2.75 meters, what proportion of the bars fail to meet the
specifications?
Exercise 19.
Let the continuous random variable X denote the current measured in a thin copper wire in
milliamperes. Assume that the range of X is [0, 20 mA], and assume that the probability density function of X
is f(x) = 0.05, for 0  x  20. What is the probability that a measurement of current is between 5 and 10
milliamperes?
(Example 4.9, “Probability and Statistics for Engineers” by Douglas and Montgomery)
Exercise 20.
Let X (in mm) be the thickness of washers a machine turns out. Assume that X has density f(x) = kx,
if 1.9 < x < 2.1 and 0 otherwise. Find k. What is the probability that a washer will have thickness between
1.95 mm and 2.05 mm? (Kreyszig 7th Ex 24.5)
Exercise 21.
The probability density function of the length of a metal rod is f(x) = 2 for 2.3 < x < 2.8 meters.
(a) If the specifications for this process are from 2.25 to 2.75 meters, what proportions of the
bars fail to meet the specifications?
(b) Assume that the probability density function is f(x) = 2 for an interval of length 0.5 meters.
Over what value the density should be centered to achieve the greatest proportion of bars
within specifications? (Douglas Montgomery)

Exercise 22.
The probability density function of the net weight in pounds of a packaged chemical herbicide is f(x)
= 2.0 for 49.75 < x < 50.25 pounds.
(a) Determine the probability that a package weights more than 50.
(b) How much chemical is obtained in 90% of all packages? (Douglas Montgomery)
The Normal Probability Distribution
A random variable X having the bell-shaped distribution is called a Normal random variable. The
probability distribution of Normal random variable is called the Normal Probability Distribution. The
probability density function of a normal random variable with mean  and standard deviations  is
1 1 x-
f(x) = f(x;  ,) = e-2 (  )2 , -   x  +
2
The normal distribution is commonly denoted by N( , 2). The distribution depends on two
parameters, the mean  and standard deviation .
How to find out the probability
b
P [a  x  b] =  f(x) dx i.e. That is the area under the curve between a and b
a
Using the Standard Normal Probability Distribution Table

Normal distribution variable take on many different units of measure: dollars, inches, pounds, time
etc. Now we talk in terms of standard units and we denote them by the symbol z. The standard normal
distribution table is organized in terms of standard units or z values. It gives the values for only half the area
under the normal curve, beginning with 0 at the mean. We can use this table for problems involving both
sides of the curve.
If Z is a normal random variable with mean 0 and standard deviation 1. Then Z is called standard
normal random variable. The probability distribution of standard normal random variable Z is called
Standard Normal Distribution. The density function is
1 -1Z2
f(z) = e2 , -   z  +
2
x-
where z= and Z  N(0, 1)

Change of variable
change the variable X to Z such that Z = (X- )/ Z is called as standard normal variable now find
the corresponding values of a and b finally find the area using table
USE OF TABLE (Tabulated Area)

A table is available for a standard normal variable. It always provides the area from 0 to a certain
positive point For other points we use the corresponding area on the positive side. For example
Examples (23)
Suppose that X is a normally distributed random variable with mean 10 and standard deviation 2.1.
(a) Find P (X  11)

(b) Find P (7.6  X  12.2)

Sol. (a)
The value x = 11 corresponds to a Z value of
x- 11-10
Z= = 2.1 = 0.48

Thus P (X  11) = P (Z  0.48)
Standard Normal curve is shown in the figure. The shaded area is the required probability-using table we
have
P (X  11) = P (Z  0.48)
` = 0.5 – 0.1844 = 0.3156
Sol. (b)
The value x1 = 7.6 and x2 = 12.2 correspond to the Z-value
x1- 7.6-10
Z1 = = 2.1 = - 1.14

x2- 12.2-10
Z2 = = 2.1 = - 1.05

P (7.6  x  12.2) = P (-1.14  Z  1.05)
= A1 + A2 = 0.3729 +0.3531
= 0.7260
Exercise 23.
Life time of an automatic dishwasher is normally distributed with mean 3.5 years and
S.D = 1.0 yrs. If a dishwasher is guaranteed for 12 months, what is the fraction of the sale will
require replacement? Mean 2,5 and SD 1.5 yrs guaranteed 6months
Exercise 24.
X denotes the height of soldiers that is normally distributed with mean = 68.22 in and vaiance = 10.8
2
in . How many soldiers in a regiment of 1000 would you expect to be over 6 feet tall?
Inverse Use of Table:
Exercise 25.
Let X  N(13.5, 1.9). Find two points such that a single observation has
95% chances to fall between them
Exercise 26.
In a normal distribution with  = 13.5 and  = 3.6. Find two points such that a single observation has
95% chance of falling between them.
Exercise 27.
The time required by an athlete to complete a race has been observed to be normally distributed with
mean 30 seconds and a standard deviation of  = 10 seconds. Find the 10th and 90th percentiles. 3rd Quartile
and 7th Decile

Examples (24)
An athlete finds that in a high jump he can clear a height of 1.68 m once in five attempts and a height
of 1.52 m nine times out of ten attempts. Assuming the heights he can clear in various jumps form a normal
distribution, estimate the mean and standard deviation of the distribution.
Solution
Let X denote the height the athlete can clear in various jumps. Then
X  N (, 2) where  and 2 are unknowns. The athlete can clear a height
of 1.68 m once in five attempts means
P[X > 1.68] = 1/5 = 0.2 and
He can clear a height of 1.52 m nine times out of ten attempts
means
P[x  1.52] = 9/10 = 0.9
 P[x  1.52] = 0.1
The probability (area) between  and x = 1.68 is 0.5 – 0.2 = 0.3
The probability (area) between  and x = 1.52 is 0.5 – 0.1 = 0.4
Using table inversely, the corresponding points are
x1 = 0.84 and x2 = - 1.28
Putting in the relation x =  + z we get
 + 0.84  = 1.68
 - 1.28 = 1.52
by solving we have
 = 1.617 and  = 0.075
Properties of a Normal Distribution
1) The function defining the normal dist. is a proper p.d.f. i.e. f(x)  0 and the total area under the
normal curve is unity.
2) The man and variance of the nomal distribution are  and 2 respectively.
3) The median and mode of the normal distribution are each equal to , the mean of the distribution.
x-
4) If X ~ N(, 2) , 2 >0 , then the random variable z = ~ N(0,1).

Examples (25)
7
If X = N(0,1), then what will be the distribution for of 2X3, X + 5 and 4X?
8
Solution
Given that E(X) = 0 and Var.(X) = 1
Let Y = 2X  3
Then E(Y) = E(2X  3) = 2E(X)  3 = 2(0) 3 = 3
and Var.(Y) = Var.( 2X  3) = 4 Var.(x)  Var (3) = 4(1)  0 = 4
Hence the distribution of 2X  3 ~ N(3, 4)
Similarly the other parts.

Normal Approximation to binomial

The normal distribution can be used as an approximation to the binomial distribution, under certain
circumstances.
If n is large and (or) p is close to ½, then
X  np
Z=
npq
is approximately a standard normal random variable.
The approximation is good for np < 5.
Examples (26)
A fair coin is tossed 20 times. Find the probability that the number of heads occurring is between 10
and 14 inclusive by using normal approximation to the binomial distribution. (Introduction to Statistical Theory
by Sher Muhammad Ch. Example 9.17)
Solution
Since np = 20(0.5) = 10 > 5, so we‟ll use the normal distribution to approximate a binomial problem.
Now  = np = 20(0.5) = 10, and
 = npq = 10 = 2.24
For normal approximation, the interval of discrete values 10  X  14 is replaced by the interval 9.5
 X  14.5.
X
We compute z values as, Z =

9.5  10
At x = 9.5, z = =  0.22
2.24
14.5  10
At x = 14.5, z = 2.24 = 2.01
Hence
P(10  X  14) = P(9.5  X  14.5) = P( 0.22 < Z < 2.01)
= 0.0871 + 0.4778 = 0.5649
Exercise 28.
A pair of dice is rolled 180 times. Use the normal approximation method to find the probability that a
total of 7 occurs (i) at least 25 times (ii) between 33 and 41 times inclusive and (iii) exactly 30 times.
(Introduction to Statistical Theory by Sher Muhammad Ch. Example 9.18)
Examples (27)
Rods are made to a nominal length of 4cm but in fact the lenmgth is a normally distributed
random variable with mean 4.01cm and standard deviation 0.03. Each rod costs 6p to make and may
be used immediately if its length lies between 3.98cm and 4.02cm. If its length is less than 3.98cm
the rod cannot be used but has a scrap value of 1p. If the length exceeds 4.02cm it can be shortened
and used at a further cost of 2p. Find the average cost per usable rod.
(HELM (VERSION 1: March 26, 2004): Workbook Level 1, 39.2: The Normal Approx. to the Binomial Distribution)
Solution
Let X denotes the length of the rod and is normally distributed with mean 4,01 cm and standard
deviation 0.03.
Cost has 2 possible values per usable rod, 6p and 8p which are associated with the probability that
the length is between 3.98cm and 4.02cm and more than 4.02cm resp.
3.98  4.01 4.02  4.01
Now P(3.98  X  4.02) = P( 0.03 <Z< 0.03 )
= P( 1 < Z < 0.333) = P(0 < Z < 1) + P(0 < Z < 0.333)
= 0.3413 + 0.1306 = 0.4718
and
P(X > 4.02) = P(Z > 0.333) = 0.5  P(0 < Z < 0.333) = 0.5  0.1306 = 0.3696
Hence
Total cost of 47.18% usable rods at the rate 6p per rod = 283.08
Total cost of 36.96% usable rods at the rate of 8p per rod = 295.6
and the cost of remaining 15.87% rods at the rate of 5p per rod = 79.35
Total Cost = 659.03
Total cost of usable rods = 84.13
Hence average cost per usable rod is = 659.03 / 84.13 = 7.82p
Examples (28)
The diameter of an electric cable is normally distributed with mean 0.8cm and variance 0.0004cm2.
(a) What is the probability that the diameter will exceed 0.81cm?
(b) The cable is considered defective if the diameter differs from the mean
by more than 0.025cm. What is the probability of obtaining a defective cable?
Solution
X ~ N(0.8, 0.0004)
(a)
0.81  0.8
P(X > 0.81) = P(Z > )
0.02
= P(Z > 0.5) = 0.5  P(0 < Z < 0.5) )
= 0.5  .1915 = 0.3085
(b)
P(X > 0.825)  P(X < 0.785) = 2 P(X > 0.825)
0.025
= 2 P(Z > 0.02 ) = 2 P(Z > 1.25)
= 2 [0.5  P(0 < Z < 1.25)]

= 2 (0.5  0.3044)
= 0.2112
Examples (29)
A machine packs sugar in what are nominally 2kg bags. However there is a variation in the actual
weight which is described by the normal distribution.
(a) Previous records indicate that the standard deviation of the distribution is 0.02 kg and the
probability that the bag is underweight is 0.01. Find the mean value of the distribution.

(b) It is hoped that an improvement to the machine will reduce the standard deviation while
allowing it to operate with the same mean value. What value standard deviation is needed to
ensure that the probability that a bag is underweight is 0.001?
Solution (a)
Given  = 0.02, and P(X < 2) = 0.01, then what is  = ?
2
i.e. P(Z < 0.02 ) = 0.01
2
 0.05  P(0 < Z < ) = 0.01
0.02
2
 P(0 < Z < 0.02 ) = 0.49
2
 = 2.33   = 2.0466
0.02
Normal Approximation to Poisson
The normal distribution can also be used to approximate the Poisson distribution for large values of , then
X  
Z=

is approximately a standard normal random variable. The approximation is good for  > 5.
Continuity Correction
To approximate a binomial or Poisson probability distribution with a normal distribution, a
continuity correction is applied as follows:
P(a  X  b) = P(a  0.5  X  b + 0.5)
Exercise 29.
The number of calls received by an office switch board per hour follow a Poisson distribution with
parameter 25. Find the probabilities that in one hour (a) there are between 23 and 26 calls inclusive (ii) more
than 30 calls, using normal approximation to the Poisson distribution.
(Introduction to Statistical Theory by Sher Muhammad Ch. Example 9.19)
Examples (30)
The annual number of earthquakes registering at least 2.5 on the Richter Scale and having an
epicenter within 40 miles of downtown Memphis follows a Poisson distribution with mean 6.5. What is the
probability that at least 9 such earthquakes will strike next year?
(An Introduction to Mathematical Statistics and Application, by Richard J. Larsen and Morris L. Marx.)
Solution
We can, of course use the Poisson distribution to calculate the exact probability. Using the Poisson
table with λ = 6.5, we get:
P(Y ≥ 9) = 1−P(Y ≤ 8) = 1− 0.792 = 0.208
Now, let's use the normal approximation to the Poisson to calculate an approximate probability. First,
we have to make a continuity correction. Doing so, we get:
P(Y ≥ 9) = P(Y > 8.5)

Once we've made the continuity correction, the calculation again reduces to a normal probability
calculation:
8.5  6.5
P(Y ≥ 9) = P(Y > 8.5) = P(Z > ) = P(Z > 0.78) = 0.218
6.5
So, in summary, we used the Poisson distribution to determine the probability that Y is at least 9 is
exactly 0.208, and we used the normal distribution to determine the probability that Y is at least 9 is
approximately 0.218. Not too shabby of an approximation!
Exercise 30.
On the basis past experience, automobile inspectors in Pennsylvania have noticed that 5 percent of
all cars coming in for their annual inspection fail to pass. Using the normal approximation to the binomial,
find the probability that between 7 and 18 of the next 200 cars to enter the Lancaster inspection station will
fail the inspection.

Table for Standard Normal Probabilities

Z .00 .01 .02 .03 .04 .05 .06 .07 .08 .09
.0 .0000 .0040 .0080 .0120 .0160 .0199 .0239 .0279 .0319 .0359
.1 .0398 .0438 .0478 .0517 .0557 .0596 .0636 .0675 .0714 .0753
.2 .0793 .0832 .0871 .0910 .0948 .0987 .1026 .1064 .1103 .1141
.3 .1179 .1217 .1255 .1393 .1331 .1368 .1406 .1443 .1480 .1517
.4 .1554 .1591 .1628 .1664 .1700 .1736 .1772 .1808 .1844 .1879
.5 .1915 .1950 .1985 .2019 .2054 .2088 .2123 .2157 .2190 .2224
.6 .2257 .2291 .2324 .2357 .2489 .2422 .2454 .2586 .2517 .2549
.7 .2580 .2611 .2642 .2673 .2703 .2734 .2764 .2794 .2823 .2852
.8 .2881 .2910 .2939 .2967 .2995 .3023 .3051 .3078 .3106 .3133
.9 .3159 .3186 .3212 .3238 .3264 .3289 .3315 .3340 .3365 .3389
1.0 .3413 .3438 .3461 .3485 .3508 .3531 .3554 .3577 .3599 .3621
1.1 .3643 .3665 .3686 .3708 .3729 .3749 .3770 .3790 .3810 .3830
1.2 .3849 .3869 .3888 .3907 .3925 .3944 .3962 .3980 .3997 .4015
1.3 .4032 .4049 .4066 .4082 .4099 .4115 .4131 .4147 .4162 .4177
1.4 .4192 .4207 .4222 .4236 .4251 .4265 .4279 .4292 .4306 .4319
1.5 .4332 .4345 .4357 .4370 .4382 .4394 .4406 .4418 .4429 .4441
1.6 .4452 .4463 .4474 .4484 .4495 .4505 .4515 .4525 .4535 .4545
1.7 .4554 .4564 .4573 .4582 .4591 .4599 .4608 .4616 .4625 .4633
1.8 .4641 .4649 .4656 .4664 .4671 .4678 .4686 .4693 .4699 .4706
1.9 .4713 .4719 .4726 .4732 .4738 .4744 .4750 .4756 .4761 .4767
2.0 .4772 .4778 .4783 .4788 .4793 .4798 .4803 .4808 .4812 .4817
2.1 .4821 .4826 .4830 .4834 .4838 .4842 .4846 .4850 .4854 .4857
2.2 .4861 .4864 .4868 .4871 .4875 .4878 .4881 .4884 .4887 .4890
2.3 .4893 .4896 .4898 .4901 .4904 .4906 .4909 .4932 .4913 .4916
2.4 .4918 .4920 .4922 .4925 .4927 .4929 .4931 .4949 .4934 .4936
2.5 .4938 .4940 .4941 .4943 .4945 .4946 .4948 .4962 .4951 .4952
2.6 .4953 .4955 .4956 .4957 .4959 .4960 .4961 .4972 .4963 .4964
2.7 .4965 .4966 .4967 .4968 .4969 .4970 .4971 .4979 .4973 .4974
2.8 .4974 .4975 .4976 .4977 .4977 .4978 .4979 .4985 .4980 .4981
2.9 .4981 .4982 .4982 .4983 .4984 .4984 .4985 .4989 .4986 .4986
3.0 .4987 .4987 .4987 .4988 .4988 .4989 .4989 .4992 .4990 .4990
3.1 .4990 .4991 .4991 .4991 .4992 .4992 .4992 .4995 .4993 .4993

05 ES Random Variables and Their Probability Distributions

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

05 ES Random Variables and Their Probability Distributions

Transféré par

Droits d'auteur :

Formats disponibles

05 ES Random Variables and their Probability Distributions (1)

Random Variables and their Probability Distributions

4th Semester MME, 2017

The probability distribution is

4th Semester MME, 2017

4th Semester MME, 2017

Number of Floods Number of A plot of probability mass function.

4th Semester MME, 2017

4th Semester MME, 2017

4th Semester MME, 2017

Weighted Average = E(X) = 0¼ + 1½ + 2¼ = 1

4th Semester MME, 2017

Which is a probability mass function for a discrete uniform random variable

4th Semester MME, 2017

Now we discuss some well known probability distributions

(1) Bernoulli Distribution

However, there are  x  tuples with x successes and n − x failures in n trials.

4th Semester MME, 2017

= 0.qn + 1.  1  p1qn-1 + 2.  2  p2qn-2 + …+ npn

= np [qn-1 +  1  pqn-2 +  2  p2qn-3 + …+ pn-1

4th Semester MME, 2017

E[X (X - 1)] =  x (x – 1)  x  pxqn-x

The term (n – x) may be written as [(n – 2) – (x - 2)]

Thus  = E(X ) – [E(X)] = E[X (X - 1)] + E(X) - [E(X)]2

when n = 5, p = 0.3 and y = 3, Thus

b. The probability of observing at least three wells containing impurity A is

4th Semester MME, 2017

Since we found p(3) in part a. we have

4th Semester MME, 2017

Mean and Variance Using mgf

= n (q + et p)n-1 et p + n(n-1) (q + et p)n-2 (et p)2

Now Var(X) = 2 = 2 - 21 = n p + n2 p2  n p2  (np)2

4th Semester MME, 2017

The characteristics of a Poisson random variable are listed below:

4th Semester MME, 2017

In other words, if X is a Binomial random variable such that

4th Semester MME, 2017

4th Semester MME, 2017

4th Semester MME, 2017

Thus, P(y 2) = 1 – P(y  1) = 1 – [P(0) + P(1)]

4th Semester MME, 2017

4th Semester MME, 2017

CONTINUOUS RANDOM VARIABLES AND THEIR PROBABILITY DISTRIBUTIONS

Expected Value of a Continuous Random Variable

Mean and Variance:

4th Semester MME, 2017

The variance of X is;

P(X < 10) =  f(x) dx =  0.05 dx = 0.5

P[5 < X < 20] =  f(x) dx =  0.05 dx = 0.5

4th Semester MME, 2017

Using the Standard Normal Probability Distribution Table

USE OF TABLE (Tabulated Area)

4th Semester MME, 2017

(b) Find P (7.6  X  12.2)

4th Semester MME, 2017

4th Semester MME, 2017

Normal Approximation to binomial

= 2 [0.5  P(0 < Z < 1.25)]

4th Semester MME, 2017

4th Semester MME, 2017

4th Semester MME, 2017

Table for Standard Normal Probabilities

4th Semester MME, 2017

Vous aimerez peut-être aussi