MODULE 1: RANDOM VARIABLE

PROBABILITY BASICS
▪Random Experiments: Experiment whose outcome is predictable but we can’t be sure
enough to say which outcome
▪Sample Space, 𝑆: Set of all possible outcome
▪Event: 𝐸 ⊂ 𝑆
▪Mutually Exclusive Events: Consider 𝐸1 and 𝐸2 be two events which cannot occur
together. 𝐸1 ∩ 𝐸2 = ∅
PROBABILITY
• Axiom 1: 0 ≤ 𝑃 𝐸 ≤ 1 for any event, 𝐸. 𝑃 𝐸 = 1: sure event, 𝑃 𝐸 = 0:
uncertain
•Axiom 2: 𝑃 𝑆 = 1
•Axiom 3: For any sequence of pairwise mutually exclusive events 𝐸1 , 𝐸2 , …
∞ ∞

𝑃 ራ 𝐸𝑖 = ෍ 𝑃(𝐸𝑖 )
𝑖=1 𝑖=1
Remark
1. 𝑃 𝐸 𝐶 = 1 − 𝑃 𝐸
2. 𝑃 ∅ = 0
CONDITIONAL PROBABILITIES
Probability of the occurrence of 𝐸 with respect other event 𝐹 has happened
𝑃 𝐸∩𝐹
𝑃 𝐸/𝐹 =
𝑃(𝐹)

INDEPENDENT EVENTS
Two event 𝐸 and 𝐹 are independent events if,
𝑃 𝐸 ∩ 𝐹 = 𝑃 𝐸 . 𝑃(𝐹)
The occurrence of 𝐸 doesn’t affect the occurrence of 𝐹

MUTUALLY EXCLUSIVE AND EXHAUSTIVE EVENTS

1. 𝐸𝑖 ∩ 𝐸𝑗 = ∅ for 𝑖 ≠ 𝑗
2. ‫=𝑖𝑛ڂ‬1 𝐸𝑖 = 𝑆
TOTAL PROBABILITY THEOREM
𝑃 𝐵 = σ𝑛𝑖=1 𝑃 𝐵/𝐸𝑖 𝑃(𝐸𝑖 )
Proof:
𝑃 𝐵∩𝐸𝑖
σ𝑛𝑖 𝑃 𝐵ൗ𝐸𝑖 𝑃 𝐸𝑖 = σ𝑛𝑖=1 𝑃(𝐸𝑖 )
𝑃 𝐸𝑖

σ𝑛𝑖 𝑃 𝐵ൗ𝐸𝑖 𝑃 𝐸𝑖 = σ𝑛𝑖=1 𝑃(𝐵 ∩ 𝐸𝑖 ) = 𝑃(𝐵)

[𝐸𝑖 are mutually exclusive and exhaustive]
BAYE’S THEOREM
𝐸𝑖 𝑃 𝐵ൗ𝐸 𝑃(𝐸𝑖 )
𝑖
𝑃 ൗ𝐵 =
σ𝑛 𝐵
𝑖=1 𝑃 ൗ𝐸𝑖 𝑃(𝐸𝑖 )

Proof:
𝐸𝑖 𝑃 𝐸𝑖 ∩𝐵 𝑃 𝐵ൗ𝐸 𝑃(𝐸𝑖 ) 𝑃 𝐵ൗ𝐸 𝑃(𝐸𝑖 )
𝑖 𝑖
𝑃 ൗ𝐵 = = = (using total probability theorem)
𝑃(𝐵) 𝑃(𝐵) σ𝑛 𝐵
𝑖=1 𝑃 ൗ𝐸𝑖 𝑃(𝐸𝑖 )
PROBLEM
1) A factory has three machines A, B, and C producing bulbs.
Machine A produces 40% of the total bulbs, with a defect rate of 2%.
Machine B produces 35% of the total bulbs, with a defect rate of 3%.
Machine C produces 25% of the total bulbs, with a defect rate of 4%.
Question:
(a) If one bulb is selected at random, what is the probability that the bulb is
defective?
(b) What is the probability that the defective bulb was produced by
(i) Machine A
(ii) Machine B
(iii) Machine C
PROBLEM2
RANDOM VARIABLE
A mapping ′𝑋′ from sample space, 𝑆 to a real number, ℝ. (i.e) It is a real valued
function defined on 𝑆.
𝑋: 𝑆 → ℝ
Eg. Tossing a coin thrice
𝑆 = 𝐻𝐻𝐻, 𝐻𝐻𝑇, 𝐻𝑇𝐻, 𝐻𝑇𝑇, 𝑇𝐻𝐻, 𝑇𝐻𝑇, 𝑇𝑇𝐻, 𝑇𝑇𝑇
𝑋 ←No. of heads obtained= {0, 1, 2, 3}
Types:
▪ Discrete Random Variable: 𝑅𝑥 is discrete
▪Continuous Random Variable: 𝑅𝑥 is continuous
DISCRETE RANDOM VARIABLE
A random variable I said to be discrete if it assume only a finite or countably infinite
value of 𝑋; that is the range space 𝑅 contains a finite or countable infinite points.
Let 𝑥1 , 𝑥2 , … be possible values of a discrete random variable 𝑋. Then 𝑃(𝑥𝑖 ) is
called the probability mass function of the discrete random variable 𝑋 if
i. 𝑃 𝑥𝑖 ≥ 0 for 𝑖 = 1,2, …
ii. σ𝑖 𝑃 𝑥𝑖 = 1
CONTINUOUS RANDOM VARIABLE
A random variable is said to be continuous if it can assume all points on the real line.
Let us now define the probability density function of a continuous random variable.
A function 𝑓 is said to be the probability density function (pdf) of a continuous
random variable 𝑋 if the following conditions are satisfied.
i. 𝑓 𝑥 ≥ 0 for all 𝑥
∞
ii. ‫׬‬−∞ 𝑓 𝑥 𝑑𝑥 = 1
CUMULATIVE DISTRIBUTIVE FUNCTION
1. Let 𝑋 be a random variable discrete or continuous. Then the function defined by
𝐹 𝑥 = 𝑃 𝑋 < 𝑥 is called cumulative distribution function.
• If 𝑋 is a discrete random variable then 𝐹 𝑥 = σ𝑥𝑖 <𝑥 𝑝(𝑥𝑖 )
𝑥
•If X is a continuous random variable then 𝐹 𝑥 = ‫׬‬−∞ 𝑓(𝑥) 𝑑𝑥
2. Every density function 𝑓 𝑥 satisfies the condition 𝐹 ∞ = 1, 𝐹 −∞ = 0
3. Also, for every real 𝑎 and 𝑏 where 𝑎 < 𝑏
𝑏
𝑃 𝑎 ≤ 𝑥 ≤ 𝑏 = න 𝑓 𝑥 𝑑𝑥
𝑎
Also if the function 𝑓(𝑥) is continuous at some point 𝑥, then 𝐹 ′ 𝑥 = 𝑓(𝑥)
PROPERTIES OF CDF:
• If 𝑥1 ≥ 𝑥2 , then 𝐹 𝑥1 ≥ 𝐹(𝑥2 ) (It is a non-decreasing function)
• lim 𝐹 𝑥 = 1
𝑥→∞
• lim 𝐹 𝑥 = 0
𝑥→−∞
•𝑃 𝑎 ≤ 𝑋 ≤ 𝑏 = 𝐹 𝑏 − 𝐹 𝑎
•𝑃 𝑋 > 𝑎 = 1 − 𝑃 𝑋 ≤ 𝑎 = 1 − 𝐹(𝑎)
PROBLEM
0, 𝑥 < 0
1 1
1. Consider 𝐹𝑥 = 𝑥 + 2 , 0 ≤ 𝑥 < 2 represents CDF of a random variable 𝑋. Find
1
1, 𝑥 ≥
2
1 1
𝑃(0 ≤ 𝑥 ≤ ) and 𝑃(𝑥 > ).
4 4
Solution:
1 1 1 1 1 1
𝑃 0≤𝑥≤ =𝐹 −𝐹 0 = + − =
4 4 4 2 2 4
1 1 1 1 1
𝑃 𝑥> =1−𝐹 =1− + =
4 4 4 2 4
2. Let 𝑋 be no. of heads when a fair coin is tossed thrice, construct
PMF, find 𝐹 1.5 , 𝐹 2.9 and construct CDF.
Solution:
The PMF is
2. Let 𝑋 be no. of heads when a fair coin is tossed thrice, construct PMF, find 𝐹 1.5 , 𝐹 2.9 and construct CDF.
Solution:
The PMF is 𝑥𝑘 0 1 2 3
𝑝𝑘 1 3 3 1
8 8 8 8
1 3 1
𝐹𝑥 1.5 = 𝑃 𝑥 ≤ 1.5 = 𝑃0 + 𝑃1 = 8 + 8 = 2
1 3 3 7
𝐹 2.9 = 𝑃 𝑥 ≤ 2.9 = 𝑃0 + 𝑃1 + 𝑃2 = 8 + 8 + 8 = 8
0, 𝑥 < 0
1
8
,0 ≤ 𝑥 < 1
1 3 4
𝐹𝑥 𝑥 = 8
+ 8
= 8
,1 ≤ 𝑥 < 2
1 3 3 7
+ + = ,2 ≤ 𝑥 < 3
8 8 8 8
1, 𝑥 ≥ 3
0, 𝑥 < 0
1
,0 ≤ 𝑥 < 1
8
4
Therefore 𝐹𝑥 𝑥 = 8
,1 ≤𝑥<2
7
8
,2 ≤𝑥<3
1, 𝑥 ≥ 3
3. Given
𝑅𝑥 = {1, 2, 3, 4} and 2 𝑃 𝑋 = 1 = 3 𝑃 𝑋 = 2 =
𝑃 𝑋 = 3 = 5(𝑃(𝑋 = 4)). Construct PMF.
Solution:
Given 2𝑃1 = 3𝑃2 = 𝑃3 = 5𝑃4
𝑘 𝑘 𝑘
Let 𝑃3 = 𝑘, 𝑃1 = , 𝑃2 = , 𝑃4 =
2 3 5
We know that σ 𝑃𝑘 = 1
𝑘 𝑘 𝑘
𝑘 + + + =1
2 3 5
30
𝑘 =
61
The P.M.F of 𝑋 is
𝑋𝑘 1 2 3 4
𝑃𝑘 15/61 10/61 30/61 6/61
 0, 𝑦 < 2
0.3, 2 ≤ 𝑦 < 4
4. Given a CDF of a random variable, 𝑅𝑦 . 𝐹𝑦 𝑦 = ൞ . Find the
0.8, 4 ≤ 𝑦 < 6
1, 𝑦 ≥ 6
corresponding PMF of 𝑌.
Solution:
𝑃 𝑌 = 2 = 0.3 − 0 = 0.3
𝑃 𝑌 = 4 = 0.8 − 0.3 = 0.5
𝑃 𝑌 = 6 = 1 − 0.8 = 0.2
The PMF is
𝑦𝑘 2 4 6
𝑃(𝑌 = 𝑦𝑘 ) 0.3 0.5 0.2
 2𝑥, 0 < 𝑥 < 𝑏
5. Consider 𝑓 𝑥 = ቊ for what value of 𝑏 does 𝑓(𝑥) represent a
0, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
PDF?
Solution:
∞
𝑊. 𝐾. 𝑇 ‫׬‬−∞ 𝑓(𝑥) 𝑑𝑥 =1
𝑏
‫׬‬0 2𝑥 𝑑𝑥 =1
𝑏2 = 1
𝑏 = ±1
Therefore, 𝑏 = 1
 1
,0 <𝑥<1
3
6. Given the PDF of 𝑋 as 𝑓 𝑥 = 2
,1 ≤ 𝑥 < 2 . Find the CDF.
3
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Solution:
0, 𝑥 < 0
𝑥
‫׬‬0 𝑓(𝑢) 𝑑𝑢, 0 ≤ 𝑥 < 1
𝐹 𝑥 = 11 𝑥
‫׬‬0 3 𝑑𝑥 + ‫׬‬1 𝑓(𝑢) 𝑑𝑢, 1 ≤ 𝑥 < 2
2
‫׬‬0 𝑓 𝑢 𝑑𝑢, 𝑥 ≥ 2
𝑥 𝑥1 x
‫׬‬0 𝑓(𝑢) 𝑑𝑢 = ‫׬‬0 3 𝑑𝑢 = 3
11 𝑥2 1 2
‫׬‬0 3 𝑑𝑥 + ‫׬‬1 3 𝑑𝑢 = 3 + 3 𝑥 − 1
2 11 22 1 2
‫׬‬0 𝑓 𝑢 𝑑𝑢 = ‫׬‬0 3 𝑑𝑢 + ‫׬‬1 3 𝑑𝑢 = 3 + 3 = 1
0, 𝑥 < 0
𝑥/3, 0 ≤ 𝑥 < 1
𝐹 𝑥 = 1 2
3
+ 3 (𝑥 − 1), 1 ≤ 𝑥 < 2
1, 𝑥 ≥ 2
 0, 𝑦 < 0
7. Given 𝐹𝑦 𝑦 = ൞3𝑦 2 − 2𝑦 3 , 0 < 𝑦 < 1. Find the PDF of 𝑦.
1, 𝑦 > 1
Solution:
0, 𝑦 < 0
𝑓 𝑦 = ൞6𝑦 − 6𝑦 2 , 0 < 𝑦 < 1
0, 𝑦 > 1

6𝑦 − 6𝑦 2 , 0 < 𝑦 < 1
𝑓 𝑦 =ቊ
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
EXPECTATION AND VARIANCE OF RANDOM VARIABLES
σ𝑘 𝑥𝑘 𝑝 𝑥𝑘 , 𝑖𝑓 𝑋 − 𝑑𝑖𝑠𝑐𝑟𝑒𝑡𝑒
𝐸 𝑋 = 𝜇𝑋 = ൝ ∞
‫׬‬−∞ 𝑥𝑓 𝑥 𝑑𝑥 , 𝑖𝑓 𝑋 − 𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑜𝑢𝑠
σ𝑘 𝑥𝑘𝑛 𝑝 𝑥𝑘 , 𝑖𝑓 𝑋 − 𝑑𝑖𝑠𝑐𝑟𝑒𝑡𝑒
𝐸 𝑋 𝑛 = 𝜇𝑋𝑛 = ൝ ∞ 𝑛
‫׬‬−∞ 𝑥 𝑓 𝑥 𝑑𝑥 , 𝑖𝑓 𝑋 − 𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑜𝑢𝑠
Variance: The measure of each quantity deviates from mean
σ𝑘(𝑥𝑘 −𝜇)2 𝑝 𝑥𝑘 , 𝑖𝑓 𝑋 − 𝑑𝑖𝑠𝑐𝑟𝑒𝑡𝑒
Var 𝑋 = 𝜎𝑋 = ቐ ∞
‫׬‬−∞(𝑥𝑘 −𝜇)2 𝑓 𝑥 𝑑𝑥 , 𝑖𝑓 𝑋 − 𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑜𝑢𝑠
2
𝑉𝑎𝑟 𝑋 = 𝐸 𝑋2 − 𝐸 𝑋
 1
8. Given 𝑓 𝑥 = , −1 < 𝑥 < 2. Find E(X) and 𝑉𝑎𝑟 (𝑋).
3
Solution:
2 2 1
𝐸 𝑋 = ‫׬‬−1 𝑥 𝑓 𝑥 𝑑𝑥 = ‫׬‬−1 𝑥 𝑑𝑥
3
1
𝐸 𝑋 =
2
2 2 1
𝐸 𝑋2 = ‫׬‬−1 𝑥 3 𝑑𝑥 = 1
1 3
𝑉𝑎𝑟 𝑥 = 𝐸 𝑋 2 − 𝐸 𝑋 2
=1− =
4 4
9. Find the 𝐸 𝑋 and 𝑉𝑎𝑟(𝑋) for the following PMF.
𝑋𝑘 1 2 3 4
𝑝𝑘 15/61 10/61 30/61 6/61

Solution:
𝐸 𝑋 = σ 𝑥𝑘 𝑝𝑘
15 10 30 6 149
𝐸 𝑋 =1 +2 +3 +4 =
61 61 61 61 61
15 10 30 6 421
𝐸 𝑋2 =1 +22 +32 + 42 =
61 61 61 61 61
421 149 2 3480
𝑉𝑎𝑟 𝑋 = 𝐸 𝑋2 −𝐸 𝑋 2 = − = = 0.9352
61 61 3721
MOMENTS
𝑟 𝑡ℎ Moment about origin:
The 𝑟 𝑡ℎ moment of a random variable 𝑋 is defined by 𝐸 𝑋 𝑟 and it is denoted by
𝜇𝑟 ′
𝜇𝑟′ = 𝐸(𝑋 𝑟 )
𝑟 𝑡ℎ Moment about point, 𝑐:
𝐸 (𝑋 − 𝑐)𝑟 is called the 𝑟 𝑡ℎ moment about the point c.
σ 𝑥𝑖 − 𝑐 𝑟 𝑝 𝑥𝑖 , 𝑋 − 𝑑𝑖𝑠𝑐𝑟𝑒𝑡𝑒
𝑟
𝐸 𝑋−𝑐 =൝ ∞
‫׬‬−∞ 𝑥𝑖 − 𝑐 𝑟 𝑓 𝑥 𝑑𝑥 , 𝑋 − 𝐶𝑜𝑛𝑡𝑖𝑛𝑢𝑜𝑢𝑠
CENTRAL MOMENTS IN TERMS OF MOMENTS
ABOUT ORIGIN
The 𝑟 𝑡ℎ central moment is given by
𝑟
𝜇𝑟 = 𝐸 𝑋 − 𝜇
𝜇𝑟 = 𝐸[𝑋 𝑟 − 𝑟𝑐1 𝑋 𝑟−1 𝜇 + 𝑟𝑐2 𝑋 𝑟−2 𝜇2 + ⋯ + −1 𝑘 𝑟𝑐𝑘 𝑋 𝑟−𝑘 𝜇𝑘 + ⋯ + −1 𝑟 𝜇𝑟 ]
′ 2 ′ 𝑘 𝑟
𝜇𝑟 = 𝜇𝑟′ − 𝑟𝑐1 𝜇𝑟−1 𝜇1′ + 𝑟𝑐2 𝜇𝑟−2
′
𝜇1′ + ⋯ + −1 𝑘 𝑟𝑐𝑘 𝜇𝑟−𝑘 𝜇1′ + ⋯ + −1 𝑟 𝜇1′
[∵ 𝜇 = 𝜇1′ , 𝜇0′ = 1]
In particular we have:
𝜇1 = 𝜇1′ − 𝜇0′ 𝜇1′ = 0
2
𝜇2 = 𝜇2′ − 𝜇1′
3
𝜇3 = 𝜇3′ − 3𝜇2′ 𝜇1′ + 2𝜇1′
2 4
𝜇4 = 𝜇4′ − 4𝜇3′ 𝜇1′ + 6𝜇2′ 𝜇1′ − 3𝜇1′
MOMENT GENERATING FUNCTION(MGF):
σ𝑖 𝑒 𝑥𝑖 𝑡 𝑝 𝑥𝑖 , 𝑋 − 𝑑𝑖𝑠𝑐𝑟𝑒𝑡𝑒
𝑀𝑋 𝑡 = 𝐸 𝑒 𝑡𝑥 = ൝ 𝑥𝑡
‫𝑥𝑑 𝑥 𝑓 𝑒 ׬‬, 𝑋 − 𝐶𝑜𝑛𝑡𝑖𝑛𝑢𝑜𝑢𝑠
Properties of MGF:
1. nth moment of R.V X can be obtained from 𝑀𝑋 (𝑡)
∞ 𝑡𝑛𝐸 𝑋𝑛
𝑀𝑋 𝑡 = σ𝑛=0
𝑛!
𝑛 𝑑 𝑛 𝑀𝑋 𝑡
2. 𝐸 𝑋 = |𝑡 = 0
𝑑𝑡 𝑛
 ∞ 𝑡𝑛𝐸 𝑋𝑛
1. nth moment of R.V X can be obtained from 𝑀𝑋 𝑡 , 𝑀𝑋 𝑡 = σ𝑛=0
𝑛!
Proof
𝑀𝑋 𝑡 = 𝐸 𝑒 𝑡𝑥
𝑡𝑥 𝑡 2𝑥2 𝑡𝑛𝑥𝑛
𝑀𝑋 𝑡 = 𝐸 1 + + +⋯+ +⋯
1! 2! 𝑛!
𝑡 𝑡2 2 𝑡𝑛
𝑀𝑋 𝑡 = 𝐸 1 + 𝐸 𝑥 + 𝐸 𝑋 +⋯+ 𝐸 𝑋𝑛 + ⋯
1! 2! 𝑛!
𝑡 𝑡2 𝑡𝑛
𝑀𝑋 𝑡 = 1 + 𝐸 𝑥 + 𝐸 𝑋2 +⋯+ 𝐸 𝑋𝑛 + ⋯
1! 2! 𝑛!
[∵ 𝐸 𝑋 satisfies linear property,
𝐸 𝑐 = 𝑐, 𝐸 𝑐𝑋 = 𝑐𝐸 𝑋 , 𝐸 𝑋1 + 𝑋2 = 𝐸 𝑋1 + 𝐸(𝑋2 )]
∞ 𝑡𝑛𝐸 𝑋𝑛
𝑀𝑋 𝑡 = σ𝑛=0
𝑛!
 𝑛 𝑑 𝑛 𝑀𝑋 𝑡
2. 𝐸 𝑋 = |𝑡 = 0
𝑑𝑡 𝑛

Proof:
𝑡 𝑡2 𝑡𝑛
𝑀𝑋 𝑡 = 1 + 𝐸 𝑋 + 𝐸 𝑋2 + ⋯+ 𝐸 𝑋𝑛 + ⋯
1! 2! 𝑛!
𝑑 1 2𝑡 2 𝑛𝑡 𝑛−1
𝑀𝑋 𝑡 = 𝐸 𝑋 + 𝐸 𝑋 + ⋯+ 𝐸 𝑋𝑛 + ⋯
𝑑𝑡 1! 2! 𝑛!
𝑑
𝑀𝑋 𝑡 𝑡 = 0 = 𝐸(𝑋)
𝑑𝑡

Similarly
𝑑2 2 𝑛(𝑛−1)𝑡 𝑛−2
𝑀𝑋 𝑡 = 𝐸 𝑋2 + ⋯+ 𝐸 𝑋𝑛 + ⋯
𝑑𝑡 2 2! 𝑛!
𝑑2
𝑀𝑋 𝑡 𝑡 = 0 = 𝐸(𝑋 2 )
𝑑𝑡 2

In general
𝑑 𝑛 𝑀𝑋 𝑡
𝐸 𝑋𝑛 = |𝑡 = 0
𝑑𝑡 𝑛
10. Given the 𝑛𝑡ℎ moment of R.V. X as 𝐸 𝑋 𝑛 = 𝑛 + 1 ! 2𝑛 . Find 𝑀𝑋 (𝑡).
Solution
Given 𝐸 𝑋 𝑛 = 𝑛 + 1 ! 2𝑛
𝑡𝑛
𝑀𝑋 𝑡 = σ∞
𝑛=0 𝑛! 𝑛 + 1 ! 2𝑛
𝑡𝑛
𝑀𝑋 𝑡 = σ∞
𝑛=0 𝑛! 𝑛 + 1 . 𝑛! 2𝑛 = σ∞ 𝑛 𝑛 ∞ 𝑛
𝑛=0 𝑡 𝑛 + 1 . 2 = σ𝑛=0(2𝑡) 𝑛 + 1

𝑀𝑋 𝑡 = σ∞
𝑛=0 𝑛 2𝑡
𝑛 + σ∞
𝑛=0 2𝑡
𝑛

1
Recall: σ∞ 𝑛 2
𝑛=0 𝑥 = 1 + 𝑥 + 𝑥 + ⋯ = 1−𝑥
1
𝐷 1 + 𝑥 + 𝑥 2 + ⋯ = 1 + 2𝑥 + 3𝑥 2 + ⋯ = σ 𝑛 𝑥 𝑛−1 =
1−𝑥 2

𝑀𝑋 𝑡 = 2𝑡 σ∞
𝑛=0 𝑛 2𝑡
𝑛−1
+ σ∞
𝑛=0 2𝑡
𝑛

2𝑡 1 1
𝑀𝑋 𝑡 = + =
1−2𝑡 2 1−2𝑡 1−2𝑡 2
 1
, −1 <𝑥<2
[Link] the MGF of a R.V. of X with PDF given by 𝑓 𝑥 = ൝3 .
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Solution
2 1
MGF= ‫׬‬−1 𝑒 𝑥𝑡 𝑑𝑥
3
2
1 2 𝑥𝑡 1 𝑒 𝑥𝑡 1 𝑒 2𝑡 𝑒 −𝑡 1
MGF= ‫𝑥𝑑 𝑒 ׬‬ = = − = [𝑒 2𝑡 − 𝑒 −𝑡 ]
3 −1 2 𝑡 −1 2 𝑡 𝑡 2𝑡
REMARK
𝑀𝑋 𝑡 = 0 = 1
σ𝑝 𝑥 = 1
𝑀𝑋 0 = ൝
‫ = 𝑥𝑑 𝑥 𝑓 ׬‬1
 2
11. Given 𝑀𝑋 𝑡 = . Obtain the mean and variance of 𝑋.
2−𝑡
Solution
𝑑[𝑀𝑋 (𝑡)]
𝐸 𝑋 = |𝑡 =0
𝑑𝑡
𝑑 −1
𝐸 𝑋 = 2 2−𝑡 |𝑡 = 0
𝑑𝑡
2
𝐸 𝑋 = |𝑡 =0
2−𝑡 2
1
𝐸 𝑋 =
2
2 𝑑 2 [𝑀𝑋 (𝑡)]
𝐸 𝑋 = |𝑡 =0
𝑑𝑡 2
1
𝐸 𝑋2 =
2
1
𝑉𝑎𝑟 𝑋 = 𝐸 𝑋2 −𝐸 𝑋 2
=
4
 2 1 𝑛
12. P.T 𝑝 𝑛 = ,𝑛 = 0,1,2, … is valid PMF of a discrete R.V. of X.
3 3
Solution
2 1 𝑛
Given 𝑝 𝑛 = ,𝑛 = 0, 1, 2, …
3 3
2 ∞ 1 𝑛
σ∞
𝑛=0 𝑝 𝑛 = σ
3 𝑛=0 3
2 1
σ∞
𝑛=0 𝑝 𝑛 = =1
3 1−1
3

And also its probabilities lie between 0 & 1

Therefore, the given function is a valid PMF
13. A random variable 𝑋 has the PDF
𝑘 𝑥 − 1 ,1 ≤ 𝑥 ≤ 2
𝑓𝑋 𝑥 = ቐ 𝑘 3 − 𝑥 , 2 ≤ 𝑥 ≤ 3
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
(a) Find the value of ‘𝑘’
(b) Find the CDF and hence determine 𝑃 1 ≤ 𝑋 ≤ 2
Solution
∞
a) ‫׬‬−∞ 𝑓𝑋 𝑥 𝑑𝑥 = 1
=> 𝑘 = 1
𝑥 − 1 ,1 ≤ 𝑥 ≤ 2
b) 𝑓𝑋 𝑥 = ቐ 3 − 𝑥 , 2 ≤ 𝑥 ≤ 3
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
 ‫𝑥‬ ‫‪𝑥2‬‬ ‫‪1‬‬
‫= 𝑥 𝑋𝐹‬ ‫𝑢(∞‪−‬׬‬ ‫‪− 1)𝑑𝑢 = − 𝑥 +‬‬
‫‪2‬‬ ‫‪2‬‬
‫‪2‬‬ ‫𝑥‬
‫𝑢 ‪1‬׬ = 𝑥 𝑋𝐹‬ ‫𝑢𝑑 𝑢 ‪2 3 −‬׬ ‪− 1 𝑑𝑢 +‬‬
‫‪6𝑥−𝑥 2 −7‬‬
‫= 𝑥 𝑋𝐹‬
‫‪2‬‬
‫‪0, 𝑥 < 1‬‬
‫)‪(𝑥 2 −2𝑥+1‬‬
‫‪,1‬‬ ‫‪≤𝑥≤2‬‬
‫‪2‬‬
‫= 𝑥 𝑋𝐹‬
‫)‪(6𝑥−𝑥 2 −7‬‬
‫‪,2‬‬ ‫‪≤𝑥≤3‬‬
‫‪2‬‬
‫‪1, 𝑥 > 3‬‬
‫‪1‬‬
‫= ‪𝑃 1≤𝑋 ≤2 =𝐹 2 −𝐹 1‬‬
‫‪2‬‬
 0, 𝑥 < 0
1
14. Can the function, 𝐹 𝑥 = 𝑥, 0 ≤ 𝑥 ≤
2 define a CDF of any random variable.
1
1, 𝑥 >
2
Solution
0, 𝑥 < 0
1
𝐹 𝑥 = 𝑥, 0 ≤ 𝑥 ≤
2
1
1, 𝑥 >
2

0, 𝑥 < 0
1
𝑓 𝑥 = 1, 0 ≤ 𝑥 ≤
2
1
1, 𝑥 >
2
1
𝑓 𝑥 = 1, 0 ≤ 𝑥 ≤
2
1
1
Since, ‫׬‬0 𝑓 𝑥 𝑑𝑥 = ≠ 1
2
2
The given function is not a valid CDF
15. The PDF of a random variable 𝑋 is 𝑓𝑋 𝑥 = 𝑎 + 𝑏𝑥 2 , 0 ≤ 𝑥 ≤ 1 and
𝐸 𝑥 = 3/5. Find ‘a’ & ‘b’
Solution
Given
𝑓𝑋 𝑥 = 𝑎 + 𝑏𝑋 2 , 0 ≤ 𝑋 ≤ 1
3
𝐸 𝑋 =
5
1 1
‫׬‬0 𝑓𝑋 𝑥 𝑑𝑥 = ‫׬‬0 (𝑎 + 𝑏𝑥 2 )𝑑𝑥 = 1
1
𝑏𝑥 3
⇒ 𝑎𝑥 + =1
3 0
𝑏 3
⇒ [𝑎 + ] =
3 5
⇒ 3𝑎 + 𝑏 = 3-------(1)
𝐸 𝑋 = 3/5
1
⇒ ‫׬‬0 𝑥𝑓 𝑥 𝑑𝑥 = 3/5
1
⇒ ‫׬‬0 𝑥(𝑎 + 𝑏𝑥 2 )𝑑𝑥 = 3/5
1
⇒ ‫׬‬0 (𝑎𝑥 + 𝑏𝑥 3 )𝑑𝑥 = 3/5
12
⇒ 2𝑎 + 𝑏 = ----(2)
5
Solving (1) and (2) we get
3 6
𝑎 = and 𝑏 =
5 5
CONDITIONAL PROBABILITY
Probability of the occurrence of 𝐸 with respect other event 𝐹 has happened
𝑃 𝐸∩𝐹
𝑃 𝐸/𝐹 =
𝑃(𝐹)
PROBLEM
1. Data on the readership of a certain magazine, the probability of male readers
under 35 is 0.40 and over 35 is 0.20. If the probability of readers under 35 is 0.70.
Calculate the probability that a randomly selected male subscriber is under 35 years
of age.
Solution:
Let us define the following events:
A: Reader of the magazine is a male
B: Reader of the magazine is over 35 years of age
Then in usual notation, we have 𝑃 𝐴 ∩ 𝐵 = 0.20, 𝑃 𝐴 ∩ 𝐵ത = 0.40, 𝑃 𝐵ത = 0.70,
𝑃 𝐵 = 0.30
𝑃 𝐴 ∩ 𝐵ത 0.40 2
ത
𝑃 𝐵 Τ𝐴 = = =
𝑃(𝐴) 0.60 3
𝑃 𝐴 = 𝑃 𝐴 ∩ 𝐵 + 𝑃 𝐴 ∩ 𝐵ത = 0.20 + 0.40 = 0.60
2. From a city population, the probability of selecting a male or a smoker is 7/10.
the probability of a male smoker is 2/5 and the probability a male, if a smoker is
already selected is 2/3. Find the probability of selecting a smoker, if a male is first
selected.
Solution:
Let A: A male is selected
B: A smoker is selected
7 2 2
Given: 𝑃 𝐴 ∪ 𝐵 = ,𝑃 𝐴 ∩ 𝐵 = , 𝑃 𝐴Τ𝐵 =
10 5 3
2
𝑃 𝐴∩𝐵 5 3
𝑃 𝐴Τ𝐵 = ⇒𝑃 𝐵 = 2 =
𝑃(𝐵) 5
3
1
𝑃 𝐴 =𝑃 𝐴∪𝐵 −𝑃 𝐵 +𝑃 𝐴∩𝐵 =
2
2
𝑃 𝐴∩𝐵 5 4
Find 𝑃 𝐵 Τ𝐴 = = 1 =
𝑃(𝐴) 5
2
TWO DIMENSIONAL RANDOM VARIABLES
•Let 𝑆 be the sample space associated with a random experiment 𝐸. Let 𝑋 = 𝑋(𝑠)
and 𝑌 = 𝑌 𝑠 be two functions each assigning a real number to each outcomes 𝑠 ∈
𝑆. Then 𝑋, 𝑌 is called a two-dimensional random variables.
•If the possible values of (𝑋, 𝑌) are finite or countably infinite. (𝑋, 𝑌) is called a two-
dimension discrete random variables.
•When 𝑋, 𝑌 is a two-dimensional discrete RV the possible values of 𝑋, 𝑌 may be
represented as 𝑥𝑖 , 𝑦𝑖 , 𝑖 = 1, 2, … , 𝑚 … ; 𝑗 = 1,2, … , 𝑛, …
•If (𝑋, 𝑌) can assume all values in a specified region 𝑅 in the 𝑥𝑦-plane, (𝑋, 𝑌) is
called a two-dimensional continuous random variables.
PROBABILITY FUNCTION OF (𝑋, 𝑌)
If (𝑋, 𝑌) is a two-dimensional discrete RV such that 𝑃 𝑋 = 𝑥𝑖 , 𝑌 = 𝑦𝑗 = 𝑝𝑖𝑗 , then 𝑝𝑖𝑗
is called probability mass function provided the following conditions are satisfied
1. 𝑝𝑖𝑗 ≥ 0, for all 𝑖 and 𝑗
2. σ𝑗 σ𝑖 𝑝𝑖𝑗 = 1
The set of triples 𝑥𝑖 , 𝑦𝑗 , 𝑝𝑖𝑗 , 𝑖 = 1,2, … , 𝑚, … ; 𝑗 = 1,2, … , 𝑛, … is called the joint
probability distribution of (𝑋, 𝑌)
JOINT PROBABILITY DENSITY FUNCTION
If 𝑋, 𝑌 is a two-dimensional continuous R.V such that
𝑑𝑥 𝑑𝑥 𝑑𝑦 𝑑𝑦
𝑃 𝑥− ≤𝑋≤𝑥+ 𝑎𝑛𝑑 𝑦 − ≤ 𝑌 ≤ 𝑦 + = 𝑓 𝑥, 𝑦 𝑑𝑥 𝑑𝑦, then 𝑓(𝑥, 𝑦)
2 2 2 2
is called the joint pdf of (𝑋, 𝑌), provided 𝑓(𝑥, 𝑦) satisfies the following conditions
1. 𝑓 𝑥, 𝑦 ≥ 0, for all (𝑥, 𝑦) ∈ 𝑅 where 𝑅 − 𝑟𝑎𝑛𝑔𝑒 𝑠𝑝𝑎𝑐𝑒
2. ‫𝑥 𝑓 𝑅׬ ׬‬, 𝑦 𝑑𝑥 𝑑𝑦 = 1
Note: If 𝐷 is a subspace of the range space 𝑅, then
𝑑 𝑏
𝑃 𝑎 ≤ 𝑋 ≤ 𝑏, 𝑐 ≤ 𝑌 ≤ 𝑑 = ‫𝑥 𝑓 𝑎׬ 𝑐׬‬, 𝑦 𝑑𝑥 𝑑𝑦
CUMULATIVE DISTRIBUTION FUNCTION
If (𝑋, 𝑌) is a two-dimensional R.V, then 𝐹 𝑥, 𝑦 = 𝑃 𝑋 ≤ 𝑥 𝑎𝑛𝑑 𝑌 ≤ 𝑦 is called the cdf of 𝑋, 𝑌
In discrete case, 𝐹 𝑥, 𝑦 = σ 𝑗 σ 𝑖 𝑝𝑖𝑗
𝑦𝑗 ≤𝑦 𝑥𝑖 ≤𝑥
𝑦 𝑥
In continuous case, 𝐹 𝑥, 𝑦 = ‫׬‬−∞ ‫׬‬−∞ 𝑓 𝑥, 𝑦 𝑑𝑥 𝑑𝑦
Properties:
1. 𝐹 −∞, 𝑦 = 0 = 𝐹(𝑥, −∞) and 𝐹 ∞, ∞ = 1
2. 𝑃 𝑎 < 𝑋 < 𝑏, 𝑌 ≤ 𝑦 = 𝐹 𝑏, 𝑦 − 𝐹 𝑎, 𝑦
3. 𝑃 𝑋 ≤ 𝑥, 𝑐 < 𝑌 < 𝑑 = 𝐹 𝑥, 𝑑 − 𝐹(𝑥, 𝑐)
4. 𝑃 𝑎 ≤ 𝑥 ≤ 𝑏, 𝑐 ≤ 𝑦 ≤ 𝑑 = 𝐹 𝑏, 𝑑 − 𝐹 𝑎, 𝑑 − 𝐹 𝑏, 𝑐 + 𝐹 𝑎, 𝑐
𝜕2 𝐹
5. At points of continuity of 𝑓(𝑥, 𝑦), = 𝑓(𝑥, 𝑦)
𝜕𝑥 𝜕𝑦
MARGINAL PROBABILITY DISTRIBUTION
If 𝑋 is discrete case:
𝑃 𝑋 = 𝑥𝑖 = σ𝑗 𝑝𝑖𝑗 is called the marginal probability function of 𝑋.
𝑃 𝑌 = 𝑦𝑗 = σ𝑖 𝑝𝑖𝑗 is called the marginal probability function of 𝑌.
If 𝑋 is continuous case:
∞
𝑓𝑋 𝑥 = ‫׬‬−∞ 𝑓 𝑥, 𝑦 𝑑𝑦 is called the marginal density of 𝑋
∞
𝑓𝑌 𝑦 = ‫׬‬−∞ 𝑓 𝑥, 𝑦 𝑑𝑥 is called the marginal density of 𝑌
CONDITIONAL PROBABILITY DISTRIBUTION
If 𝑋 is discrete case:
𝑃 𝑋=𝑥𝑖 ,𝑌=𝑦𝑗 𝑝𝑖𝑗
•𝑃 𝑋 = 𝑥𝑖 Τ𝑌 = 𝑦𝑗 = = is called the conditional probability function of 𝑋,
𝑃(𝑌=𝑦𝑗 ) 𝑝∗𝑗
given that 𝑌 = 𝑦𝑗
𝑝𝑖𝑗
•The collection of pairs 𝑥𝑖 , , 𝑖 = 1, 2, … is called the conditional probability distribution of
𝑝∗𝑗
𝑋 given 𝑌 = 𝑦𝑗
𝑃 𝑋=𝑥𝑖 ,𝑌=𝑦𝑗 𝑝𝑖𝑗
•Similarly, 𝑃 𝑌 = 𝑦𝑗 Τ𝑋 = 𝑥𝑖 = = is called the conditional probability function
𝑃(𝑋=𝑥𝑖 ) 𝑝𝑖∗
of 𝑌, given that X= 𝑥𝑖
𝑝𝑖𝑗
•The collection of pairs , 𝑦𝑗 , 𝑖 = 1, 2, … is called the conditional probability distribution of
𝑝𝑖∗
𝑌 given X= 𝑥𝑖
CONDITIONAL PROBABILITY DISTRIBUTION
If 𝑋 is continuous case:
𝑓 𝑥,𝑦
• is called the conditional density function of 𝑋, given 𝑌 and it is denoted by 𝑓(𝑥 ∕ 𝑦)
𝑓𝑌 (𝑦)
𝑓 𝑥,𝑦
• is called the conditional density function of 𝑌, given 𝑋 and it is denoted by 𝑓(𝑦 ∕ 𝑥)
𝑓𝑋 (𝑥)
PROBLEMS
1. Three balls are drawn at random without replacement from a box containing 2 white, 3 red
and 4 black balls. If 𝑋 denotes the number of white balls drawn and 𝑌 denotes the number of
red balls drawn, find the joint probability distribution of 𝑋, 𝑌 .
Solution:
As there are only 2 white balls in the box, X can take the value 0, 1 and 2.
Similarly, as there are only 3 red balls in the box, Y can take the value 0, 1, 2 and 3.
𝑃 𝑋 = 0, 𝑌 = 0 = 𝑃(𝑑𝑟𝑎𝑤𝑖𝑛𝑔 3 𝑏𝑎𝑙𝑙𝑠 𝑛𝑜𝑛𝑒 𝑜𝑓 𝑤ℎ𝑖𝑐ℎ 𝑖𝑠 𝑤ℎ𝑖𝑡𝑒 𝑜𝑟 𝑟𝑒𝑑)
4𝐶3 1
𝑃 𝑋 = 0, 𝑌 = 0 = 𝑃 𝐴𝑙𝑙 𝑡ℎ𝑟𝑒𝑒 𝑏𝑎𝑙𝑙𝑠 𝑎𝑟𝑒 𝑏𝑙𝑎𝑐𝑘 = =
9𝐶3 21
3𝐶1 ×4𝐶2 3
𝑃 𝑋 = 0, 𝑌 = 1 = 𝑃 𝐷𝑟𝑎𝑤𝑖𝑛𝑔 1 𝑟𝑒𝑑 𝑏𝑎𝑙𝑙 𝑎𝑛𝑑 2 𝑏𝑙𝑎𝑐𝑘 𝑏𝑎𝑙𝑙𝑠 = =
9𝐶3 14
3𝐶2 ×4𝐶1 1
𝑃 𝑋 = 0, 𝑌 = 2 = 𝑃 𝐷𝑟𝑎𝑤𝑖𝑛𝑔 2 𝑟𝑒𝑑 𝑏𝑎𝑙𝑙 𝑎𝑛𝑑 1 𝑏𝑙𝑎𝑐𝑘 𝑏𝑎𝑙𝑙 = =
9𝐶3 7
3𝐶3 1
𝑃 𝑋 = 0, 𝑌 = 3 = 𝑃 𝐷𝑟𝑎𝑤𝑖𝑛𝑔 3 𝑟𝑒𝑑 𝑏𝑎𝑙𝑙𝑠 = =
9𝐶3 84
 2𝐶1 ×4𝐶2 1
𝑃 𝑋 = 1, 𝑌 = 0 = 𝑃 𝑑𝑟𝑎𝑤𝑖𝑛𝑔 1 𝑤ℎ𝑖𝑡𝑒 𝑏𝑎𝑙𝑙 𝑎𝑛𝑑 2 𝑏𝑙𝑎𝑐𝑘 𝑏𝑎𝑙𝑙𝑠 = 9𝐶3
=7
2𝐶1 ×3𝐶1 ×4𝐶1 2
𝑃 𝑋 = 1, 𝑌 = 1 = 𝑃 𝑑𝑟𝑎𝑤𝑖𝑛𝑔 1 𝑤ℎ𝑖𝑡𝑒 𝑏𝑎𝑙𝑙, 1 𝑟𝑒𝑑 𝑏𝑎𝑙𝑙 𝑎𝑛𝑑 1 𝑏𝑙𝑎𝑐𝑘 𝑏𝑎𝑙𝑙 = 9𝐶3
=7
2𝐶1 ×3𝐶2 1
𝑃 𝑋 = 1, 𝑌 = 2 = 𝑃 𝐷𝑟𝑎𝑤𝑖𝑛𝑔 1 𝑤ℎ𝑖𝑡𝑒 𝑏𝑎𝑙𝑙 𝑎𝑛𝑑 2 𝑟𝑒𝑑 𝑏𝑎𝑙𝑙𝑠 = =
9𝐶3 14
3𝐶2 ×4𝐶1 1
𝑃 𝑋 = 1, 𝑌 = 2 = 𝑃 𝐷𝑟𝑎𝑤𝑖𝑛𝑔 2 𝑟𝑒𝑑 𝑏𝑎𝑙𝑙 𝑎𝑛𝑑 1 𝑏𝑙𝑎𝑐𝑘 𝑏𝑎𝑙𝑙 = =
9𝐶3 7

𝑃 𝑋 = 1, 𝑌 = 3 = 0 (Since only 3 balls are drawn)

2𝐶2 ×4𝐶1 1
𝑃 𝑋 = 2, 𝑌 = 0 = 𝑃 𝑑𝑟𝑎𝑤𝑖𝑛𝑔 2 𝑤ℎ𝑖𝑡𝑒 𝑏𝑎𝑙𝑙𝑠 𝑎𝑛𝑑 1 𝑏𝑙𝑎𝑐𝑘 𝑏𝑎𝑙𝑙 = 9𝐶3
=7
2𝐶1 ×3𝐶1 ×4𝐶1 2
𝑃 𝑋 = 2, 𝑌 = 1 = 𝑃 𝑑𝑟𝑎𝑤𝑖𝑛𝑔 1 𝑤ℎ𝑖𝑡𝑒 𝑏𝑎𝑙𝑙, 1 𝑟𝑒𝑑 𝑏𝑎𝑙𝑙 𝑎𝑛𝑑 1 𝑏𝑙𝑎𝑐𝑘 𝑏𝑎𝑙𝑙 = =7
9𝐶3
2𝐶1 ×3𝐶2 1
𝑃 𝑋 = 2, 𝑌 = 2 = 𝑃 𝐷𝑟𝑎𝑤𝑖𝑛𝑔 1 𝑤ℎ𝑖𝑡𝑒 𝑏𝑎𝑙𝑙 𝑎𝑛𝑑 2 𝑟𝑒𝑑 𝑏𝑎𝑙𝑙𝑠 = = 14
9𝐶3
3𝐶2 ×4𝐶1 1
𝑃 𝑋 = 2, 𝑌 = 3 = 𝑃 𝐷𝑟𝑎𝑤𝑖𝑛𝑔 2 𝑟𝑒𝑑 𝑏𝑎𝑙𝑙 𝑎𝑛𝑑 1 𝑏𝑙𝑎𝑐𝑘 𝑏𝑎𝑙𝑙 = =7
9𝐶3
Joint probability distribution of (𝑋, 𝑌) is given as
X Y
0 1 2 3
0 1 3 1 1
21 14 7 84
1 1 2 1 0
7 7 14
2 1 1 0 0
21 28
2. For the bivariate probability distribution of (𝑋, 𝑌) is given below, find 𝑃 𝑋 ≤ 1 ,
𝑃 𝑌 ≤ 3 , 𝑃 𝑋 ≤ 1, 𝑌 ≤ 3 , 𝑃 𝑋 ≤ 1Τ𝑌 ≤ 3 , 𝑃 𝑌 ≤ 3Τ𝑋 ≤ 1 .
X/Y 1 2 3 4 5 6
0 0 0 1/32 2/32 2/32 3/32
1 1/16 1/16 1/8 1/8 1/8 1/8
2 1/32 1/32 1/64 1/64 0 2/64
Solution
1 5 7
𝑃 𝑋 ≤1 =𝑃 𝑋 =0 +𝑃 𝑋 =1 = + =
4 8 8
X/Y 1 2 3 4 5 6 Marginal
𝑋
0 0 0 1/32 2/32 2/32 3/32 ¼
1 1/16 1/16 1/8 1/8 1/8 1/8 5/8
2 1/32 1/32 1/64 1/64 0 2/64 2/16
Marginal 𝑌 3/32 3/32 11/64 13/64 6/32 1/4

3 3 11 23
𝑃 𝑌 ≤ 3 = 𝑃 𝑌 = 1 + 𝑃 𝑌 = 2 + 𝑃(𝑌 = 3) = + + =
32 32 64 64
𝑃 𝑋 ≤ 1, 𝑌 ≤ 3 = σ3𝑗=1 𝑃 𝑋 = 0, 𝑌 = 𝑗 + σ3𝑗=1 𝑃(𝑋 = 1, 𝑌 = 𝑗)
1 1 1 1 9
⇒ 𝑃 𝑋 ≤ 1, 𝑌 ≤ 3 = 0 + 0 + + + + =
32 16 16 8 32
9
𝑃 𝑋≤1,𝑌≤3 32 9
𝑃 𝑋 ≤ 1Τ𝑌 ≤ 3 = = 7 =
𝑃(𝑌≤3) 28
8
3. The joint probability mass function of (𝑋, 𝑌) is given by 𝑝 𝑥, 𝑦 =
𝑘 2𝑥 + 3𝑦 , 𝑥 = 0, 1,2, ; 𝑦 = 1, 2, 3 . Find all the marginal and
conditional probability distributions of X, given 𝑌 = 1.
Solution:
The joint probability mass function (𝑋, 𝑌) is given below
X Y
1 2 3
0 3𝑘 6𝑘 9𝑘
1 5𝑘 8𝑘 11𝑘
2 7𝑘 10𝑘 13𝑘

σ3𝑗=1 σ2𝑖=0 𝑝 𝑥𝑖 , 𝑥𝑗 = 1
The sum of all probabilities in the table is equal to 1.
72𝑘 = 1
1
=> 𝑘 =
72
Marginal probability distribution:
X Y
1 2 3 Marginal 𝑋
0 3 6 9 18
72 72 72 72
1 5 8 11 24
72 72 72 72
2 7 10 13 30
72 72 72 72
Marginal 𝑌 15 24 33 1
27 72 72
𝑃 𝑋=𝑖,𝑌=1
Conditional distribution of 𝑋, given 𝑌 = 1, is given by 𝑃 𝑋 = 𝑖 Τ𝑌 = 1 = , 𝑖 = 0,1,2
𝑃(𝑌=1)
𝑋=𝑖 𝑃(𝑋 = 𝑖 Τ𝑌 = 1)
0 𝑃 𝑋 = 0, 𝑌 = 1 1
=
𝑃(𝑌 = 1) 5
1 1/3
2 7/15
Total = 1
INDEPENDENT RV’S
•If (𝑋, 𝑌) is a two-dimensional discrete RV such that 𝑃 𝑋 = 𝑥𝑖 Τ𝑌 = 𝑦𝑗 =
𝑃(𝑋 = 𝑥𝑖 ) or 𝑝𝑖𝑗 = 𝑝𝑖∗ × 𝑝∗𝑗 for all 𝑖, 𝑗 then 𝑋 and 𝑌 are said to be
independent RV’s
•If (𝑋, 𝑌) is a two-dimensional continuous RV such that 𝑓 𝑥, 𝑦 = 𝑓𝑋 (𝑥) ×
𝑓𝑌 (𝑦), then 𝑋 and 𝑌 are said to be independent RV’s
4.A machine is used for a particular job in the forenoon and for a
different job in the afternoon. The joint probability distribution of
𝑋, 𝑌 , where 𝑋 and 𝑌 represent the number of times the machine
breaks down in the forenoon and in the afternoon respectively, is
given in the following table. Examine 𝑋 and 𝑌 are independent R.V’s.

X Y
0 1 2
0 0.1 0.04 0.06
1 0.2 0.08 0.12
2 0.2 0.08 0.12
Solution:
If 𝑋 and 𝑌 are independent, then 𝑝𝑖𝑗 = 𝑝𝑖∗ × 𝑝𝑗∗ for all 𝑖 and 𝑗.

X Y
0 1 2 Marginal 𝑋
0 0.1 0.04 0.06 0.2
1 0.2 0.08 0.12 0.4
2 0.2 0.08 0.12 0.4
Marginal 𝑌 0.5 0.2 0.3 1
Here 𝑝0∗ = 0.2, 𝑝1∗ = 0.4, 𝑝2∗ = 0.4, 𝑝∗0 = 0.5, 𝑝∗1 = 0.2, 𝑝∗2 = 0.3
𝑝0∗ × 𝑝∗0 = 0.2 × 0.5 = 0.1 = 𝑝00
𝑝0∗ × 𝑝∗1 = 0.2 × 0.2 = 0.04 = 𝑝01
Similarly we can verity for other elements
Since it is satisfied
𝑋 and 𝑌 are independent
 𝑥2
5. The joint pdf of a two-dimensional RV 𝑋, 𝑌 is given by 𝑓 𝑥, 𝑦 = 𝑥𝑦 2 + ,
8
1 1
0 ≤ 𝑥 ≤ 2, 0 ≤ 𝑦 ≤ 1. compute 𝑃 𝑋 > 1 , 𝑃 𝑌 < , 𝑃 𝑋 > 1, 𝑌 < , 𝑃 ቀ𝑌 <
1 2 2
Τ𝑋 > 1 ቁ and 𝑃(𝑋 < 𝑌).
2
Solution:
2 𝑥2
𝑓 𝑥, 𝑦 = 𝑥𝑦 + ,0 ≤ 𝑥 ≤ 2; 0 ≤ 𝑦 ≤ 1
8
1 1 2 𝑥2 19
𝑖 𝑃 𝑋>1 = ‫׬‬0 ‫>𝑥׬‬1 𝑓 𝑥, 𝑦 𝑑𝑥 𝑑𝑦 = ‫׬‬0 ‫׬‬1 𝑥𝑦 2 + 𝑑𝑥 𝑑𝑦 =
8 24
1 2 1/2 2 𝑥2 1
(ii) 𝑃 𝑌 < = ‫׬ <𝑦׬‬0 𝑓
1 𝑥, 𝑦 𝑑𝑥 𝑑𝑦 = ‫׬‬0 ‫׬‬1 𝑥𝑦 2 + 𝑑𝑥 𝑑𝑦 =
2 2 8 4
1 1/2 2 𝑥2 5
(iii) 𝑃 𝑋 > 1, 𝑌 < = ‫׬‬0 ‫׬‬1 𝑥𝑦 2 + 𝑑𝑥 𝑑𝑦 =
2 8 24
1 5
1 𝑃 𝑌<2,𝑋>1 5
(iv) 𝑃 𝑌 < Τ𝑋 >1 = = 24
19 =
2 𝑃 𝑋>1 19
24
1 𝑦 𝑥2 53
(v) 𝑃 𝑋 < 𝑌 = ‫׬‬0 ‫׬‬0 𝑥𝑦 2 + 𝑑𝑥𝑑𝑦 =
8 480
 2 +𝑦 2 )
7. The joint pdf of the RV 𝑋, 𝑌 is given by 𝑓 𝑥, 𝑦 = 𝑘𝑥𝑦𝑒 −(𝑥 , 𝑥 > 0, 𝑦 > 0.
Find the value of 𝑘 and prove that 𝑋 and 𝑌 are independent.
Solution:
∞ ∞ 2 +𝑦 2 )
The joint p.d.f ‫׬‬0 ‫׬‬0 𝑘𝑥𝑦𝑒 −(𝑥 𝑑𝑥 𝑑𝑦 = 1
∞ 2 ∞ 2
⇒ 𝑘 ‫׬‬0 𝑦𝑒 −𝑦 𝑑𝑦 . ‫׬‬0 𝑥𝑒 −𝑥 𝑑𝑥 = 1
𝑘
⇒ =1
4
⇒𝑘=4
∞ 2 +𝑦 2 ) 2
Now, 𝑓𝑋 𝑥 = ‫׬‬0 4𝑥𝑦𝑒 −(𝑥 𝑑𝑦 = 2𝑥𝑒 −𝑥 , 𝑥 > 0
∞ −(𝑥 2 +𝑦 2 ) −𝑦 2
Similarly, 𝑓𝑌 𝑦 = ‫׬‬0 4𝑥𝑦𝑒 𝑑𝑥 = 2𝑦𝑒 ,𝑥 >0
−(𝑥 2 +𝑦 2 )
𝑓𝑋 𝑥 × 𝑓𝑌 𝑦 = 4𝑥𝑦𝑒 = 𝑓(𝑥, 𝑦)
Therefore RV 𝑋 and 𝑌 are independent
8. Given 𝑓𝑋𝑌 𝑥, 𝑦 = 𝑐𝑥 𝑥 − 𝑦 , 0 < 𝑥 < 2, −𝑥 < 𝑦 < 𝑥 and 0 elsewhere, (a) evaluate 𝑐, (b) find 𝑓𝑋 (𝑥), (c)
𝑓𝑌Τ𝑋 (𝑦/𝑥) and (d) 𝑓𝑌 (𝑦).
Solution:
(a) W.K.T ‫𝑥(𝑓 𝑥׬ 𝑦׬‬, 𝑦) 𝑑𝑥 𝑑𝑦 = 1
2 𝑥
‫׬‬0 ‫׬‬−𝑥 𝑐𝑥 (𝑥 − 𝑦) 𝑑𝑦 𝑑𝑥 = 1
⇒ 8𝑐 = 1
1
⇒𝑐=8
𝑥 𝑥 𝑥−𝑦 𝑥3
(b) 𝑓𝑋 𝑥 = ‫׬‬−𝑥 8 𝑑𝑦 = 4 , 0 <𝑥<2
𝑓 𝑥,𝑦 𝑥−𝑦
(c) 𝑓𝑌Τ𝑋 𝑦Τ𝑥 = = 2 , −𝑥 <𝑦<𝑥
𝑓𝑋 (𝑥) 2𝑥
2 1
(d) 𝑓𝑌 𝑦 = ‫׬‬−𝑦 𝑥 𝑥 − 𝑦 𝑑𝑥 , 𝑖𝑛 −2<𝑦<0
8
21
𝑓𝑌 𝑦 = ‫ 𝑦׬‬8 𝑥 𝑥 − 𝑦 𝑑𝑥 , 𝑖𝑛 0 < 𝑦 < 2

1 𝑦 5
− + 𝑦 3 , −2 < 𝑦 < 0
𝑓𝑌 𝑦 = ൞31 4𝑦 481
3
3
− 4
+ 48
𝑦 ,0 < 𝑦 < 2
COVARIANCE
•Variance = 𝐸 𝑋 2 − 𝐸 𝑋 2

•Covariance = 𝑐𝑜𝑣 𝑥, 𝑦 = 𝐸 𝑋𝑌 − 𝐸 𝑋 𝐸(𝑌)

•If 𝑋 and 𝑌 are independent [Link] then 𝐸 𝑋𝑌 = 𝐸 𝑋 𝐸 𝑌 ⇒ 𝐶𝑜𝑣 𝑋, 𝑌 = 0
•If 𝐶𝑜𝑣 𝑋, 𝑌 > 0, then as 𝑋 increases, 𝑌 also increases
•If 𝐶𝑜𝑣 𝑋, 𝑌 < 0, then as 𝑋 increases, 𝑌 tends to decrease
Suppose 𝑋 and 𝑌 are continuous random variables with joint p.d.f
𝑓 𝑥, 𝑦 = 3𝑥, 𝑤ℎ𝑒𝑟𝑒 0 ≤ 𝑦 ≤ 𝑥 ≤ 1
Find the covariance of 𝑋 and 𝑌.
Solution:
1 𝑥 3
𝐸 𝑋𝑌 = ‫𝑥 𝑓𝑦𝑥 ׬ ׬‬, 𝑦 𝑑𝑦𝑑𝑥 = ‫׬‬0 ‫׬‬0 𝑥𝑦 3𝑥 𝑑𝑦𝑑𝑥 =
10
𝐸 𝑋 = 3/4
𝐸 𝑌 = 3/8
3
𝐶𝑂𝑉 𝑋, 𝑌 = 𝐸 𝑋𝑌 − 𝐸 𝑋 𝐸 𝑌 =
160
 𝑘𝑥, 0 < 𝑥 < 1, 0 < 𝑦 < 1
A joint density function is given by 𝑓 𝑥, 𝑦 = ቊ
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
FUNCTION OF A RANDOM VARIABLE
Let 𝑋be a random variable defined on a sample space 𝑆.
If 𝑔 ⋅ is a real-valued function, then
𝑌=𝑔 𝑋
is called a function of the random variable 𝑋.

Discrete Random Variable Case Continuous Random Variable Case

Method: Method:
If 𝑋is discrete with pmf 𝑃 𝑋 = 𝑥 ,then: If 𝑋is continuous with pdf 𝑓𝑋 𝑥 ,and
[Link] all possible values of 𝑌 = 𝑔 𝑋 𝑌=𝑔 𝑋
[Link] Then the pdf of 𝑌is obtained using CDF method or
𝑃 𝑌=𝑦 =𝑃 𝑔 𝑋 =𝑦 transformation formula:
𝑑𝑥
𝑓𝑌 𝑦 = 𝑓𝑋 𝑥
𝑑𝑦
where 𝑥 = 𝑔−1 𝑦