12-Sep-19

PROBABILITY

1 2

PROBABILITY
A quantitative measure of uncertainty
A measure of the strength of belief in the occurrence of an uncertain
event
A measure of the degree of chance or likelihood of occurrence of an
uncertain event
Measured by a number between 0 and 1 (or between 0% and 100%)

2 types of probability
Objective or Classical Probability Subjective Probability
 based on equally-likely events  based on personal beliefs,
 based on long-run relative experiences, prejudices,
frequency of events intuition – personal judgment
 not based on personal beliefs  different for all observers
 is the same for all observers (subjective)
(objective)

3 4
 12-Sep-19

Equally-likely Probabilities Basic Rules for Probability

(Hypothetical or Ideal Experiments)
For example:
Roll a dice
Six possible outcomes 𝑆 = 1,2,3,4,5,6
If each is equally-likely, the probability of each is = 0.1667 =
16.67%
1
𝑃 𝑒 =
𝑛(𝑆)

Set A: outcome is an even number

1 1 1 1
𝑃 𝐴 =𝑃 2 +𝑃 4 +𝑃 6 = + + =
6 6 6 2

𝑛(𝐴)
𝑃 𝐴 = 𝑃 𝑒 =
𝑛(𝑆)

5 6

Combinatorial Concepts Tree diagram of Combinatorial concepts

7 8
 12-Sep-19

Order is not important - Combinations Order is important - Permutations

9 10

Conditional Probabilities
Factorials
- a measure of the probability of an event (A) occurring
given that another event (B) has occurred.
- Probability of A given B, noted by 𝑷 𝑨|𝑩
𝑷(𝑨∩𝑩)
𝑷 𝑨|𝑩 = , where 𝑃(𝐵) ≠ 0
𝑷(𝑩)

So, 𝑷 𝑨 ∩ 𝑩 = 𝑷 𝑨|𝑩 . 𝑷 𝑩 = 𝑷 𝑩|𝑨 . 𝑷 𝑨

If events A and D are statistically independent:

𝑷 𝑨|𝑫 = 𝑷 𝑨
so, 𝑃 𝐴 ∩ 𝐷 = 𝑃 𝐴 . 𝑃(𝐷)
𝑷 𝑫|𝑨 = 𝑷 𝑫
Example: Picking a card of Ace and a card of Heart in a card deck
is statistically independent, prove it.

11 12
 12-Sep-19

Product Rules for Independent events The Law of Total Probabilities

- a fundamental rule relating marginal probabilities to
conditional probabilities
𝑷 𝑨 = 𝑷 𝑨 ∩ 𝑩 + 𝑷(𝑨 ∩ 𝑩)

In terms of conditional probabilities:

𝑃 𝐴 = 𝑃(𝐴 ∩ 𝐵) + 𝑃(𝐴 ∩ 𝐵)
= 𝑃 𝐴|𝐵 . 𝑃 𝐵 + 𝑃 𝐴|𝐵 . 𝑃(𝐵)

More generally, where 𝐵 make up a partition:

𝑃 𝐴 = 𝑃(𝐴 ∩ 𝐵 ) = 𝑃 𝐴|𝐵 . 𝑃(𝐵 )

13 14

Bayes’ Theorem Contingency Table and Joint Probability Table

- Bayes’ theorem enables you, knowing just a little more than - A joint probability table is similar to a contingency table,
the probability of A given B, to find the probability of B given A except that it has probabilities in place of frequencies
- The row totals and column totals are called marginal
Example: calculate 𝑃 𝐵|𝐴 based on 𝑃 𝐴|𝐵 probabilities.
𝑷 𝑩|𝑨 =
( ∩ ) Example: Two groups of individuals were tested the effect of inoculation
( ) in Cholera attack. One group had inoculated medicines, the other had no
( ∩ ) medicines. Calculate the probability that an individual has no cholera’
*apply the law of total probability =
∩ ( ∩ ) attacked given that individual got inoculated medicines.
𝑷 𝑨|𝑩 .𝑷(𝑩) Attacked Unattacked Total
*apply the conditional probability =
𝑷 𝑨|𝑩 .𝑷 𝑩 𝑷 𝑨|𝑩 .𝑷(𝑩) Inoculated 11 89 100
Uninoculated 21 79 100
Total 32 168 200
- Bayes’ theorem extended - given a partition of events 𝐵 , 𝐵 , … , 𝐵 : Attacked Unattacked Total
( ∩ ) ( ∩ ) ( | ). ( ) Inoculated 0.055 0.445 0.50
𝑃 𝐵 |𝐴 = = ∑ ( ∩ )
=∑ | . ( )
( ) Uninoculated 0.105 0.395 0.50
Total 0.160 0.840 1.00

15 16
 12-Sep-19

TYPES of RANDOM VARIABLES

A discrete random variable:
 has a countable number of hossible
values
 Has discrete jumps (or gaps) between  Binomial distribution
successive values  Poisson distribution
 has measurable probability associated
with individual values
 counts
A continuous random variable:
DISTRIBUTION OF  has an uncountably infinite number
 (Standard) Normal distribution
of possible values
PROBABILITY  moves continuously from value to  t-distribution
value  F-distribution
 has no measurable probability  χ -distribution
associated with each value
 measures

18 19

Discrete Random Variables Example

- The probability distribution of a discrete random variable X
must satisfy the following two conditions:
1. 𝑷 𝒙 ≥ 𝟎 for all values of 𝑥
2. ∑𝒂𝒍𝒍 𝒙 𝑷 𝒙 = 𝟏
[Corollary: 0 ≤ 𝑃(𝑋) ≤ 1]

- The cumulative distribution function, 𝑭 𝒙 , of a discrete

random variable X is:
𝑭 𝒙 = 𝑷 𝑿 ≤ 𝒙 = ∑𝒂𝒍𝒍 𝒊 𝒙 𝑷(𝒊)

20 21
 12-Sep-19

Example Example

22 23

Binomial Distribution Example

24 25
 12-Sep-19

Poisson Distribution Poisson Distribution

27 28

Example

29