Probability

Distributions
Week 1
The classical probability concept
• If there are n equally likely possibilities, of which
one must occur and s are regarded as favorable, or
𝑠
as a success, then the probability of a success is:
𝑛
Basic rules of probability
1. Probabilities are real numbers between 0 and 1
2. If an event is certain to occur, its probability is 1,
and if an event is certain not to occur, its
probability is 0
3. If two events are mutually exclusive, the
probability that one or the other will occur equals
the sum of their probabilities
4. The sum of the probabilities that an event will
occur and that it will not occur is equal to 1
Random variable
• A quantity that can take on the different values of a
given set with specified probabilities
• Contains an element of chance
• Strictly - a random variable is a function
• Loosely - quantities that can take on different
values depending on chance
• Classifies according to the number of values that
they can assume
• Discrete – can take on only finite number of values
• Continuous – a countable infinity of values (determined
by means of continuous measuring scales such as
thermometers, stopwatches)
Probability distributions - The probability of a discrete value occurring
discrete random variables x P(x)
0 0.1
• x is a discrete random variable which 1 0.1
can take the values 0, 1, 2, 3, 4, 5 and 6
2 0.15
• Which value is most likely to occur? 3 0.3

• Need information on the probability of 4 0.2

each value occurring 5 0.1
6 0.05
• Probability distribution for the random
variable x
P(x)
• The probabilities sum to 1 0.35
0.3
• If x is sampled 100 times then on 0.25
0.3

average 0 will occur 10 times; 1 will 0.2

occur 10 times; 2 will occur 15 times 0.15
0.2
0.15
and so on 0.1
0.1 0.1 0.1
0.05
• The probability that X will equal to 2 is
0.05
0
0 1 2 3 4 5 6
0.15, that is, x
𝑃 𝑋 = 2 = 0.15
Plotted data of the probability values
Expected values of discrete
random variables
• The mean of X or expected value
𝜇=𝐸 𝑥 = 𝑥 ∙ 𝑝(𝑥)

• The variance of X or the expected value of the squared

distance from the mean
𝜎2 = 𝐸 𝑥 − 𝜇 2
= 𝑥 − 𝜇 2 𝑝(𝑥)

• The standard deviation of a discrete random variable

𝜎 = 𝜎2
 The probability that x lies in a given subinterval
Probability density function for x P(x)
continuous random variables [0, 0.2] 0.1
• x is a continuous random variable [0.2, 0.4] 0.25
which can take any value on [0, 1]
[0.4, 0.6] 0.35
• Impossible to list all possible values [0.6, 0.8] 0.2
• Infinitely many values on [0, 1] [0.8, 1.0] 0.1

• The probability of any particular P(x)

value occurring is zero 0.4

• “What is the probability of x falling in 0.35

0.35
a sub-interval [0, 1]?” 0.3

0.25
• Dividing [0, 1] into sub-intervals and 0.25
0.2
attaching probabilities to each sub- 0.2
0.15
interval will give a probability
0.1
distribution 0.1 0.1
0.05
• The probability that x will lie 0
between 0.4 and 0.6 is 0.35, that is, 0.2 0.4 0.6 0.8 1.0
x
𝑃 0.4 ≤ 𝑋 < 0.6 = 0.35 Plotted data of the probability values
 The probability that x lies in a given subinterval
x P(x) x P(x)

[0, 0.1] 0.03 [0.5, 0.6] 0.2

• Making the sub-intervals [0.1, 0.2] 0.07 [0.6, 0.7] 0.1
smaller, a more refined
distribution is obtained [0.2, 0.3] 0.1 [0.7, 0.8] 0.1

[0.3, 0.4] 0.15 [0.8, 0.9] 0.07

[0.4, 0.5] 0.15 [0.9, 1.0] 0.03

P(x)
0.25

0.2
0.2

0.15
0.15 0.15

0.1
0.1 0.1 0.1

0.05 0.07 0.07

0.03 0.03
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
x
Plotted data of the probability values
Probability density function (pdf)

Shaded area represents 𝑃(𝑎 ≤ 𝑋 ≤ 𝑏)

𝑃 𝑎 ≤ 𝑋 ≤ 𝑏 = area above 𝑎, 𝑏 = 𝑓 𝑥 𝑑𝑥
𝑎
Expected value of a continuous
random variable
• The mean of X or expected value
𝑏

𝜇=𝐸 𝑥 = 𝑥 ∙ 𝑓 𝑥 𝑑𝑥
𝑎
• The variance of X or the expected value of the squared
distance from the mean
𝑏

𝜎2 = 𝐸 𝑥 − 𝜇 2
= 𝑥 − 𝜇 2 𝑓(𝑥 )𝑑𝑥
𝑎
• The standard deviation of a continuous random
variable
𝜎 = 𝜎2
Discrete Probability
Distributions
Binomial Distribution
• Repeated independent trials
• Assumptions about this distribution:
(a) There is a fixed number of trials, each resulting in
success or failure
(b) The probability of a success is the same for each trial
(c) The trials are all independent
• The probability of getting “x successes in n trials” or
“x successes and n - x failures in n attempts”
Binomial Distribution (cont.)
• In general, if n is the number of trials, p is the probability of
a success for each trial, and the trials are all independent,
the probability of x successes in n trials is
𝑛 𝑥
𝑃 𝑋 = 𝑥 = 𝑓(𝑥) = 𝑝 (1 − 𝑝)𝑛−𝑥
𝑥
for x = 0, 1, 2, …, or n
• 𝑃 𝑋 = 𝑥 is called the probability distribution function (pdf)
• Mean of the binomial distribution
𝜇 = 𝑛𝑝
• Standard deviation of the binomial distribution
𝜎 = 𝑛𝑝(1 − 𝑝)
Hypergeometric Distribution
• Sample without replacement from a finite population
• In general, if n objects are chosen at random from a set
consisting of a objects of type A and b objects of type
B, and the selection is without replacement, the
probability of “x of type A and n – x of type B” is
𝑎 𝑏
𝑥 𝑛−𝑥
𝑃 𝑋 = 𝑥 = 𝑓(𝑥) =
𝑎+𝑏
𝑛
for x = 0, 1, 2, …, or n
• x cannot exceed a and n – x cannot exceed b. Also, x
cannot exceed n
Hypergeometric Distribution (cont.)
• Binomial approximation to hypergeometric
distribution:
𝑛 ≤ (0.05)(𝑎 + 𝑏)
Poisson Distribution
• Counted events with no fixed upper limit
• Poisson processes involve observing discrete events
in a continuous “interval” of time, length, or space
• The variable of interest in a Poisson process is
X = the number of occurrences of the event in
an interval of length s units
Poisson Distribution (cont.)
• The probability of getting x “successes” by means
of the formula
𝑒 −𝜆 𝜆𝑥
𝑃 𝑋=𝑥 =𝑓 𝑥 =
𝑥!
for x = 0, 1, 2, 3, …
• The parameter 𝜆 of a Poisson process represents
the average number of occurrences of the event in
question per measurement unit
Poisson Distribution (cont.)
• Poisson approximation to binomial distribution
𝑒 −𝑛𝑝 𝑛𝑝 𝑥
𝑃 𝑋=𝑥 =𝑓 𝑥 =
𝑥!
for x = 0, 1, 2, 3, …
Multinomial distribution
• Generalization of the binomial distribution
• When there are more than two possible outcomes
for each trial
• The probabilities of the various outcomes remain
the same for each trial
• Trials are all independent
Multinomial distribution (cont.)
• If there are k possible outcomes for each trial and
their probabilities are 𝑝1 , 𝑝2 , …, and 𝑝𝑘
• The probability of 𝑥1 outcomes of the first kind, 𝑥2
outcomes of the second kind, … , and 𝑥𝑘 outcomes
of the kth kind in n trials is given by
𝑛! 𝑥1 𝑥2 𝑥𝑘
𝑃 𝑋=𝑥 =𝑓 𝑥 = 𝑝1 𝑝2 … 𝑝𝑘
𝑥1 ! 𝑥2 ! … 𝑥𝑘 !
Exercise 1
To pass a quality control inspection, two cellular phones are
randomly chosen from each lot of 10 cellular phones, and the
lot is passed only if neither phone is defective. Determine the
probabilities that a lot will
(a) pass the inspection when only 1 of the 10 phones is
defective
(b) pass the inspection when 2 of the 10 phones are defective

(c) pass the inspection when 3 of the 10 phones are defective

(d) pass the inspection when 4 of the 10 phones are defective

Exercise 2
A quality control engineer wants to check whether (in
accordance with specifications) 95% of the products shipped
are in perfect condition. To this end, he randomly selects 10
items from each large lot ready to be shipped and passes it
only if they are all in perfect condition; otherwise, each item
in the lot is checked. Find the probabilities that he will commit
the error of
(a) holding the lot for further inspection even though 95% of
the items are in perfect condition
(b) letting a lot pass through even though only 90% of the
items are in perfect condition
(c) letting a lot pass through even though only 80% of the
items are in perfect condition
Exercise 3
In the inspection of tin plate produced by a
continuous electrolytic process, 0.2 imperfections
are spotted per minute, on average. Find the
probabilities of spotting
(a) one imperfection in 3 minutes
(b) at least two imperfection in 5 minutes
(c) at most one imperfection in 15 minutes
Exercise 4
Among 20 defective glass bricks, 10 have cracks but
no discoloration, 7 are discolored but have no cracks,
and 3 have cracks and discoloration.
 What is the probability that among 6 bricks
chosen at random for further checks 3 will have
cracks only, 2 will only be discolored, and 1 will
have cracks as well as discoloration?
Exercise 5
In the inspection of a fabric produced in continuous
rolls, the mean number of imperfections spotted by
an inspector during a 5-minute period is 3.4.
 What are the probabilities that during the 5-
minute period an inspector will spot
(a) more than one imperfection?
(b) at most five imperfections?
 What are the probabilities that during the 10-
minute period an inspector will spot
(c) three or four imperfections?
(d) at least six imperfections?
Exercise 6
The percentage distribution What is the probability that
of money incomes in US among twelve randomly
households in 1995 was chosen households, one has
income under $14,999, three
Percentage of
Dollars
households have incomes between
$15,000 and $49,999, seven
Under 14,999 21.0
have incomes between
15,000-49,999 30.1
$50,000 and $74,999, and
50,000-74,999 34.0 one has income $75,000 and
75,000 and over 14.9 over?
Continuous Probability
Distributions
Normal Distribution
• The probability distribution function (pdf) for
normal distribution is defined by
1 1 𝑥−𝜇 2
−
𝑃 𝑋=𝑥 =𝑓 𝑥 = 𝑒 2 𝜎
𝜎 2𝜋
mean, 𝐸 𝑋 = 𝜇 and standard deviation, 𝑉 𝑋 = 𝜎 2
• Standard Normal Distribution, N(0, 1)
𝑋−𝜇
𝑍=
𝜎
mean, 𝐸 𝑋 = 0 and standard deviation, 𝑉 𝑋 = 1
The Normal Approximation to the
Binomial Distribution
• A normal distribution with
𝜇 = 𝑛𝑝 and 𝜎 = 𝑛𝑝(1 − 𝑝)
can often be used to approximate a binomial
distribution, even when n is fairly small and p
differs from 12
• A good rule of thumb is to use this approximation
only when np and np (1 - p) are both greater than
5; symbolically, when
𝑛𝑝 > 5 and 𝑛𝑝 1 − 𝑝 > 5
The Normal Approximation to the
Binomial Distribution (cont.)
• If X is a binomial random variable, then
𝑋 − 𝑛𝑝
𝑍=
𝑛𝑝(1 − 𝑝)
is approximately a standard normal random
variable, N(0,1)
The Normal Approximation to
Poisson Distributions
• If X is a Poisson random variable with 𝐸 𝑋 = 𝜆
and 𝑉 𝑋 = 𝜆, then
𝑋−𝜆
𝑍=
𝜆
is approximately a standard normal random
variable, N(0,1)
Exponential Distribution
• The random variable X that equals the distance
between successive counts of a Poisson process
with mean 𝜆 > 0 has an exponential distribution
with parameter 𝜆
• The probability density function of X is
𝑃 𝑋 = 𝑥 = 𝑓 𝑥 = 𝜆𝑒 −𝜆𝑥
• If the random variable X has an exponential
distribution with parameter 𝜆 , then its mean,
1 1
𝐸 𝑋 = and standard deviation, 𝑉 𝑋 = 2
𝜆 𝜆
Exercise 7
In an experiment to determine the amount of time
required to assemble an “easy to assemble” toy, the
assembly time was found to be a random variable
having approximately a normal distribution with
mean 27.8 minutes and standard deviation 4.0
minutes. What are the probabilities that this kind of
toy can be assembled in
(a) less than 25.0 minutes?
(b) anywhere from 26.0 to 29.6 minutes?
Exercise 8
A representative of a computer company claims that
20% of customer complaints result from the failure
of customers to follow operating instructions. If this
claim is correct, what is the probability that among
100 complaints more than 25 were caused by failure
of customers to follow operating instructions?
Exercise 9
Suppose that the number of false alarms received
weekly by the fire department of a large city follows
a Poisson distribution with mean 45. In any given
week, find the probability that there will be
(a) 50 or more false alarms;
(b) at most 44 false alarms.
Exercise 10
If on the average three trucks arrived per hour to be
unloaded at a warehouse, what are the probabilities
that the time between the arrivals of successive
trucks will be
(a) less than 5 minutes
(b) at least 45 minutes
Joint Probability
Distributions
Linear Combinations of Random
Variables
• A random variable is sometimes defined as a
function of several random variables
• Example:
• one manufactured part
• X1 and X2 denote the length and width, respectively
• 𝑌 = 2𝑋1 + 2𝑋2 is a random variable that represents the
perimeter of the part
Definition
• Given random variables 𝑋1 , 𝑋2 , 𝑋3 , … , 𝑋𝑝 and
constants 𝑐1 , 𝑐2 , 𝑐3 , … , 𝑐𝑝 . Then
𝑌 = 𝑐1 𝑋1 + 𝑐2 𝑋2 + ⋯ + 𝑐𝑝 𝑋𝑝

is a linear combination of 𝑋1 , 𝑋2 , 𝑋3 , … , 𝑋𝑝
Mean and Variance

• The mean for Y are defined as 𝐸(𝑌)

𝐸 𝑌 = 𝑐1 𝐸 𝑋1 + 𝑐2 𝐸 𝑋2 + ⋯ + 𝑐𝑝 𝐸 𝑋𝑝

• The variance for Y are defined 𝑉(𝑌)

𝑉 𝑋 = 𝑐12 𝑉 𝑋1 + 𝑐22 𝑉 𝑋2 + ⋯ + 𝑐𝑝2 𝑉(𝑋𝑝 )
Reproductive property for
independent, normal random variables
• If 𝑋1 , 𝑋2 , 𝑋3 , … , 𝑋𝑝 are independent, normal
random variables with 𝐸 𝑋𝑖 = 𝜇𝑖 and 𝑉 𝑋𝑖 = 𝜎𝑖2
for 𝑖 = 1, 2, 3, … , 𝑝. Then
𝑌 = 𝑐1 𝑋1 + 𝑐2 𝑋2 + 𝑐3 𝑋3 + ⋯ + 𝑐𝑝 𝑋𝑝
is a normal random variable with mean
𝐸 𝑌 = 𝑐1 𝜇1 + 𝑐2 𝜇2 + 𝑐3 𝜇3 + ⋯ + 𝑐𝑝 𝜇𝑝
and variance
𝑉 𝑋 = 𝑐12 𝜎12 + 𝑐22 𝜎22 + 𝑐32 𝜎32 + ⋯ + 𝑐𝑝2 𝜎𝑝2
Exercise 11
Suppose the random variables X1 and X2 denote the
length and width, respectively, of a manufactured
part. Assume E(X1) = 2 cm with standard deviation
0.1 cm and E(X2) = 5 cm with standard deviation 0.2
cm. Determine the probability that the perimeter
exceeds 14.5 cm. Find the mean and variance
perimeter of the part.
Exercise 12
The time to complete a coating process has mean 35
minutes and variance 11, while the time to rinse has
mean 8 minutes and variance 5.
Find the mean and standard deviation of the total
time to coat and rinse.