Probability and Statistics for

Data Science
Dr. Faisal Bukhari
Punjab University College of Information Technology
(PUCIT)
 Textbooks

Probability & Statistics for Engineers & Scientists,

Ninth Edition, Ronald E. Walpole, Raymond H.
Myer

Elementary Statistics: Picturing the World, 6th

Edition, Ron Larson and Betsy Farber

Elementary Statistics, 13th Edition, Mario F. Triola

Dr. Faisal Bukhari, PUCIT, PU, Lahore 2

 Reference books
 Probability Demystified, Allan G. Bluman

 Schaum's Outline of Probability and Statistics

 MATLAB Primer, Seventh Edition

 MATLAB Demystified by McMahon, David

Dr. Faisal Bukhari, PUCIT, PU, Lahore 3

Reference books
 Probability and Statistical Inference, Ninth Edition,
Robert V. Hogg, Elliot A. Tanis, Dale L. Zimmerman

 Probability Demystified, Allan G. Bluman

 Practical Statistics for Data Scientists: 50 Essential

Concepts, Peter Bruce and Andrew Bruce

Schaum's Outline of Probability, Second Edition,

Seymour Lipschutz, Marc Lipson

Python for Probability, Statistics, and Machine

Learning, José Unpingco
Dr. Faisal Bukhari, PUCIT, PU, Lahore 4
References
Readings for these lecture notes:

Probability & Statistics for Engineers & Scientists,

Ninth edition, Ronald E. Walpole, Raymond H. Myer

http://www.statisticshowto.com/geometric-
distribution/

https://peakmemory.me/category/forgetting-
curve/

These notes contain material from the above

resources.
Dr. Faisal Bukhari, PUCIT, PU, Lahore 5
“If you want to know what a man's like, take a good
look at how he treats his inferiors, not his equals.”

― J.K. Rowling, Harry Potter and the Goblet of Fire

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Forgetting curve

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Poisson Distribution [1]
Example: 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 fault is
a Poisson random variable with λ = 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?
PDr. Faisal Bukhari, PUCIT, PU, Lahore
Solution: Here λt=(5)(1)=5
(λt)xe−λt
P(x; λt) = , x = 0, 1, 2,. . .
x!

(a)P(X ≤ 3) = σx=3
x=0 p(x; 5)
= 0.2650

(b) P(X > 𝟏) = 1 - P(x≤ 1)

= 1 - σ𝑥=1
𝑥=0 𝑝(𝑥; 5)
= 1 - 0.0404
= 0.9596

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Poisson Distribution [2]
Example: Changes in airport procedures require
considerable planning. Arrival rates of aircraft are
important factors that must be taken into account.
Suppose small aircraft arrive at a certain airport,
according to a Poisson process, at the rate of 6 per
hour. Thus the Poisson parameter for arrivals for a
period of hours is λ = 6.

(a) What is the probability that exactly 4 small aircraft

arrive during a 1-hour period?

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Poisson Distribution [3]
(b) What is the probability that at least 4 arrive during
a 1-hour period?

(c) If we define a working day as 12 hours, what is the

probability that at least 75 small aircraft arrive during
a day?

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Poisson Distribution [3]
Here λt=(6)(1)=6
(λt)xe−λt
P(x; λt) = , x = 0, 1, 2, . . .
x!
(6)4 e−6
(a) P(X = 4) = = 0.1339
4!
(b) P(X ≥ 4) = 1 - P(x< 4) = 1- σx=3 x=0 p(x; 6) = 1 -
0.1512 = 0.8488
Here λt = (6)(12) = 72
(c) P(X ≥ 75) = 1- P(x < 75) = 1- σx=74 x=0 p(x; 72) =
0.3773.

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Poisson Distribution using Python
a) What is the probability that exactly 4 small aircraft
arrive during a 1-hour period?
from scipy.stats import poisson
mu = 6
x = 4
prob = round(poisson.pmf(x, mu), 4)
print('Probability that at least 4
arrive during a1-hour period:', prob)
#0.1339
Dr. Faisal Bukhari, PUCIT, PU, Lahore
(b) What is the probability that at least 4 arrive during
a 1-hour period?
x = [0, 1, 2, 3]
prob = 1 - round(sum(poisson.pmf(x,
mu)), 4)
print('Probability that at least 4
arrive during a 1-hour period:',
prob)
#0.8488

Dr. Faisal Bukhari, PUCIT, PU, Lahore

(c) If we define a working day as 12 hours, what is the
probability that at least 75 small aircraft arrive during a
day?
mu = 12 * 6
x = range(0, 75)
#x = list(range(0, 75))
#print(x)
prob =1 - round(sum(poisson.pmf(x,
mu)), 4)
print('Probability that at least 75
small aircraft arrive during a day:',
prob)
#0.3773
Dr. Faisal Bukhari, PUCIT, PU, Lahore
Poisson approximation
The Binomial distribution converges towards
the Poisson distribution as the number of trials goes to
infinity while the product np remains fixed. Therefore
the Poisson distribution with parameter λ= np can be
used as an approximation to b(n, p) of the binomial
distribution if n is sufficiently large and p is sufficiently
small.

According to two rules of thumb, this approximation is

good if
n ≥ 20 and p ≤ 0.05, or if n ≥ 100 and np ≤ 10.
Dr. Faisal Bukhari, PUCIT, PU, Lahore
Poisson Distribution [4]
Formula:

f(x) = (e-λ λx)/x! , x = 0, 1, 2, …

where, λ is an average rate of value, x is a Poisson
random variable and e is the base of logarithm(e =
2.718).

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Example:

Consider, in an office on average 2 customers arrived

per day. Calculate the possibilities for exactly 3
customers to be arrived on today.

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Step1: Find e-λt.
where, λt = 2 and e = 2.718, e-λt = (2.718)-2 = 0.135.

Step2: Find (λt)x

where, t =1, λt = 2 and x =3, (λt)x = 23 = 8.

Step3: Find P(x; λ)

e−λt (λt)x
P(x; λ) = , x = 0, 1, 2, …
x!
(0.135)(8)
P(3; 2) = = 0.18.
3!
Hence there are 18% possibilities for 3 customers to be
arrived today

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Geometric Distribution [1]
Suppose we have a sequence of Bernoulli trials, each
with a probability p of success and a probability q =
1-p of failure. How many trials occur before we
obtain a success?

Example
A search engine goes through a list of sites looking
for a given key phrase. Suppose the search
terminates as soon as the key phrase is found. The
number of sites visited is Geometric.

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Let the random variable X be the number of trials
needed to obtain a success. Then X has values in the
range {1,2,…}, and for k ≥1,
g(x; p) = p 𝐪𝐱−𝟏 , 𝐱 = 1, 2, 3, · · ·

Alternative form
g(x; p) = p 𝐪𝐱 , 𝐱 = 0, 1, 2, 3, · ·

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Geometric Distribution [2]
Mean = 1/p and Variance = q/p²

In the theory of probability and statistics, a Bernoulli

trial is an experiment whose outcome is random and
can be either of two possible outcomes, “success’’ and
“failure’’.

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Geometric Distribution [3]
Conditions:
An experiment consists of repeating trials until first
success.
Each trial has two possible outcomes.
A success with probability p.
A failure with probability q = 1 − p.
Repeated trials are independent.
x = number of trials to first success
x is a GEOMETRIC RANDOM VARIABLE.
g(x; p) = 𝐪𝐱−𝟏 p, 𝐱 = 1, 2, 3, · · ·

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Assumptions for the Geometric
Distribution
The three assumptions are:
There are two possible outcomes for each trial
(success or failure).

The trials are independent.

The probability of success is the same for each trial.

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Example From past experience it is known that 3% of
accounts in a large accounting population are in error.
What is the probability that 5 accounts are audited
before an account in error is found?

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Solution:
g(x; p) = 𝐪𝐱−𝟏 p, 𝐱 = 1, 2, 3, · · ·

P(X = 5) = P(1st 4 correctly stated) P(5th in error)

= g(x; p) = 𝐪𝐱−𝟏 p, 𝐱 = 1, 2, 3, · · ·
= (0.97)5−1 (0.03)
= 0.0266

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Example: In a certain manufacturing process it is
known that, on the average, 1 in every 100, items is
defective. What is the probability that the fifth item
inspected is the first defective item found?

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Solution: Using the geometric distribution with x = 5
and
p = 1/100 = 0.01, q = 0.99, we have

g(x; p) = p 𝐪𝐱−𝟏 , 𝐱 = 1, 2, 3, · · ·

g(5;0.01) = (0.01)(0.99)5−1
= 0.0096

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Python code
from scipy.stats import geom
p = 1/100
x = 5
prob = round(geom.pmf(x, p), 4)
print('The probability that the fifth
item inspected is the first defective
item found :', prob)
# 0.0096

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Example: At “busy time” a telephone exchange is very
near capacity, so callers have difficulty placing their
calls. It may be of interest to know the number of
attempts necessary in order to gain a connection.
Suppose that we let p = 0.05 be the probability of a
connection during busy time. We are interested in
knowing the probability that 5 attempts are necessary
for a successful call.

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Solution:
Using the geometric distribution with x = 5 and p =
0.05 yields

g(x; p) = p qx−1 , x = 1, 2, 3, · · ·

P(X = x) = g(5; 0.05)

= (0.05) (0.95)5−1

= 0.041.

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Python code
p = 0.05
x = 5
prob = round(geom.pmf(x, p), 4)

print('The probability that 5

attempts are necessary for a
successful call :', prob)

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Discrete Uniform Distribution [1]
If a random variable has any of n possible values that
are equally probable, then it has a discrete uniform
distribution. The probability of any outcome ki is 1/n.

A simple example of the discrete uniform distribution

is throwing a fair die. The possible values of k are 1, 2,
3, 4, 5, 6; and each time the die is thrown, the
probability of a given score is 1/6.

Dr. Faisal Bukhari, PUCIT, PU, Lahore

 Discrete Uniform Distribution [2]
Generating random numbers are the prime application
of uniform distribution. The basic random numbers are
0, 1, 2, 3, 4, 5, 6, 7, 8, 9. Each with probability equal to
1/10.

For two digit random numbers the probability of

selecting a particular random variable will be 1/100.

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Discrete Uniform Distribution [3]
If the random variable X assumes the values 𝑥1 , 𝑥1 , 𝑥2 ,
…, 𝑥𝑘 with equal probabilities, then the discrete
uniform distribution is given by

𝟏
P(x; k) = , 𝐱𝟏 , 𝐱 𝟐 , 𝐱𝟑 , …, 𝐱𝐤
𝐤

Dr. Faisal Bukhari, PUCIT, PU, Lahore

Discrete Uniform Distribution [4]
When a light bulb is selected at random from a box that
contains a 40-watt bulb, a 60-watt bulb, a 75-watt bulb,
and a 100-watt bulb, each element of the sample1
space S = {40, 60, 75, 100} occurs with probability 1/4.
Therefore, we have a uniform distribution, with
probability

𝟏
P(x; k) = , 𝐱 = 𝟒𝟎, 𝟔𝟎, 𝟕𝟓, 𝟏𝟎𝟎
𝟒

Dr. Faisal Bukhari, PUCIT, PU, Lahore