Unit -III Joint Probability Density Function

TWO DIMENSIONAL RANDOM VARIABLES Let (𝑋, 𝑌) be a two dimensional continuous R.V., then the function 𝑓(𝑥, 𝑦) is
Definitions: called the joint probability density function of the R.V. (𝑋, 𝑌), if it satisfies
Two Dimensional Random Variable the following conditions.
Let 𝑆 be the sample space. Let 𝑋 = 𝑋(𝑠) and 𝑌 = 𝑌(𝑠) be two functions each (i) 𝑓 (𝑥, 𝑦) ≥ 0, 𝑓𝑜𝑟 (𝑥, 𝑦)𝜖𝑅, Where 𝑅 is the range space.
assigning a real number to each outcome 𝑠𝜖𝑆. Then (𝑋, 𝑌) is a two (ii) ∫𝑅 ∫ 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 = 1
dimensional random variable. 𝑑 𝑏
Note: 𝑃{𝑎 ≤ 𝑋 ≤ 𝑏, 𝑐 ≤ 𝑌 ≤ 𝑑} = ∫𝑐 ∫𝑎 𝑓 (𝑥, 𝑦)𝑑𝑥𝑑𝑦
Types of Two Dimensional Random Variables
(i) Discrete Random Variable
Joint Cumulative Distribution Function
If the possible values of (𝑋, 𝑌) are finite or countably infinite, (𝑋, 𝑌) is called
If (𝑋, 𝑌) is a two dimensional R.V (discrete or continuous), then 𝐹(𝑥, 𝑦) =
a two dimensional discrete R.V. The possible values of a two dimensional
𝑃{X ≤ x , Y ≤ y} is called the c.d.fof(𝑋, 𝑌).
discrete R.V (𝑋, 𝑌) may be represented as (𝑥𝑖 , 𝑦𝑖 ) 𝑖 = 1,2, … … 𝑚 … . 𝑎𝑛𝑑 𝑗 =
In the discrete case, 𝐹 (𝑥, 𝑦) = ∑𝑗 ∑𝑖 𝑝𝑖𝑗 , 𝑦𝑗 ≤ 𝑦, 𝑥𝑖 ≤ 𝑥
1,2, … … 𝑛 … ..
𝑦 𝑥
(ii) Continuous Random Variable In the Continuous case, 𝐹 (𝑥, 𝑦) = ∫−∞ ∫−∞ 𝑓 (𝑥, 𝑦)𝑑𝑥𝑑𝑦.

If (𝑋, 𝑌) can assume all values in a specified region R in the 𝑥𝑦-plane, then Marginal Probability Distributions (or) Marginal Distributions
(𝑋, 𝑌)is called a two dimensional continuous R.V (a)Discrete Case
Let 𝑓(𝑥, 𝑦) be the joint p.d.f. of a continuous two dimensional R.V. (𝑋, 𝑌).
∞
Joint Probability Distribution (or) Probability Mass Function of (𝑿, 𝒀) The marginal density function of𝑋 is defined by 𝑓𝑋 (𝑥) = ∫−∞ 𝑓 (𝑥, 𝑦)𝑑𝑦.
∞
Let (𝑋, 𝑌)be a two dimensional discrete r.v. such that 𝑃(X = xi , Y = yj ) = 𝑝𝑖𝑗 The marginal density function of 𝑌 is defined by 𝑓𝑌 (𝑦) = ∫−∞ 𝑓 (𝑥, 𝑦)𝑑𝑥 .
, then 𝑝𝑖𝑗 is called the probability function of (𝑋, 𝑌)(joint probability (b)Continuous Case
distribution) provided the following conditions are satisfied. 𝑏 ∞ 𝑏
Note: 𝑃[𝑎 ≤ 𝑋 ≤ 𝑏] = ∫𝑎 ∫−∞ 𝑓 (𝑥, 𝑦)𝑑𝑦𝑑𝑥 = ∫𝑎 𝑓𝑥 (𝑥)𝑑𝑥
(i) 𝑝𝑖𝑗 ≥ 0 𝑓𝑜𝑟 𝑖 𝑎𝑛𝑑 𝑗 (ii) ∑𝑗 ∑𝑖 𝑝𝑖𝑗 = 1 𝑑 ∞ 𝑑
𝑃[𝑐 ≤ 𝑌 ≤ 𝑑] = ∫ ∫ 𝑓 (𝑥, 𝑦)𝑑𝑥𝑑𝑦 = ∫ 𝑓𝑦 (𝑦)𝑑𝑦
The set of triples (𝑥𝑖 , 𝑦𝑗 , 𝑝𝑖𝑗 ) 𝑖 = 1,2, … 𝑚 … 𝑎𝑛𝑑 𝑗 = 1,2, … . 𝑛 …. is called the 𝑐 −∞ 𝑐

joint probability distribution of(𝑋, 𝑌).

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 1
 Problems Based on Discrete Random Variables
Conditional Probability Distribution Problem: 1 From the following joint distribution of 𝑋 𝑎𝑛𝑑 𝑌, find the
(a)Discrete Case marginal distributions
The Conditional probability function of 𝑋 given 𝑌 = 𝑦 is defined by X
𝑃{𝑋 = 𝑥 ∩ 𝑌 = 𝑦} 𝑝𝑖𝑗 Y 0 1 2
𝑃{𝑋 = 𝑥/𝑌 = 𝑦} = =
𝑃{𝑌 = 𝑦} 𝑝∗𝑗 0 3/28 9/28 3/28
The conditional probability function of 𝑌 given 𝑋 = 𝑥 is defined by 1 3/14 3/14 0
𝑃{𝑋 = 𝑥 ∩ 𝑌 = 𝑦} 𝑝𝑖𝑗 2 1/28 0 0
𝑃{𝑌 = 𝑦/𝑋 = 𝑥} = =
𝑃{𝑋 = 𝑥} 𝑝𝑖∗
(b)Continuous Case
Solution:
The Conditional probability function of 𝑋 given 𝑌 = 𝑦 is defined by
The marginal distribution of 𝑋 𝑎𝑛𝑑 𝑌 are given below
𝑥 𝑓(𝑥, 𝑦)
𝑓( ) = , 𝑓 (𝑦) > 0 Marginal
𝑦 𝑓𝑦 (𝑦) 𝑦 X
The conditional probability function of 𝑌 given 𝑋 = 𝑥 is defined by Distribution of Y
0 1 2
𝑦 𝑓(𝑥, 𝑦) ∑ 𝑝(𝑥, 𝑦)
𝑓( ) = , 𝑓 (𝑥) > 0 Y
𝑥 𝑓𝑥 (𝑥) 𝑥 𝑥
Independent Random Variables 3 9 3 15
0
Two discrete random variables 𝑋 𝑎𝑛𝑑 𝑌 are said to be independent, if 28 28 28 28
𝑃[𝑋 = 𝑥 ∩ 𝑌 = 𝑦] = 𝑃 (𝑋 = 𝑥). 𝑃(𝑌 = 𝑦) 3 3 6
1 0
In case of continuous R.V., the random variables 𝑋 𝑎𝑛𝑑 𝑌 are said to be 14 14 14
1 1
independent, if 2 0 0
28 28
𝑓 (𝑥, 𝑦) = 𝑓𝑥 (𝑥). 𝑓𝑦 (𝑦)
Marginal
Note: Let 𝐹 (𝑥, 𝑦) be the joint probability distribution function. Then the
Distribution of X 10 15 3
joint probability density function of𝑋 𝑎𝑛𝑑 𝑌is given by 1
28 28 28
𝜕 2 𝐹(𝑥, 𝑦) ∑ 𝑝(𝑥, 𝑦)
𝑓 (𝑥, 𝑦) =
𝜕𝑥𝜕𝑦 𝑦

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 2
 Problem:2 The jointdistribution of𝑋 𝑎𝑛𝑑 𝑌 is given by X
0 1 2 ∑ 𝑝(𝑥, 𝑦)
𝑥+𝑦
𝑓(𝑥 , 𝑦) = 21
, 𝑥 = 1,2,3; 𝑦 = 1,2.Find the marginal distributions. Y 𝑥

Solution: 1 3𝑘 5𝑘 7𝑘 15𝑘
The marginal distribution of 𝑋 𝑎𝑛𝑑 𝑌 are given below 2 6𝑘 8𝑘 10𝑘 24𝑘
Marginal 3 9𝑘 11𝑘 13𝑘 33𝑘
X Distribution of
∑ 𝑝(𝑥, 𝑦) 18𝑘 24𝑘 30𝑘 72𝑘
1 2 3 Y 𝑦
Y
∑ 𝑝(𝑥, 𝑦) We know that ∑3𝑗=1 ∑2𝑖=0 𝑝(𝑥𝑖, 𝑦𝑗 ) = 1 i,e.,total probability = 1.
𝑥 1
72𝑘 = 1 ⇒ 𝑘 = .
2 3 4 9 72
1 (i)The marginal distribution of𝑋 𝑎𝑛𝑑 𝑌 are given below.
21 21 21 21
3 4 5 12 Marginal
2 X
21 21 21 21 Distribution of
Marginal 0 1 2 Y
Distribution of X 5 7 9 Y
1 ∑ 𝑝(𝑥, 𝑦)
21 21 21
∑ 𝑝(𝑥, 𝑦) 𝑥
𝑦 3 5 7 15
1 (𝑥 ∩ 𝑦) (𝑝( 𝑌) = 𝑦)
72 72 72 72
Problem: 3 If the joint distribution of 𝑋 𝑎𝑛𝑑 𝑌 is given by 𝑝(𝑥 , 𝑦) = 𝐾(2𝑥 + 6 8 10 24
2
3𝑦), 𝑦 = 1,2,3 ; 𝑥 = 0,1,2Find (i) all the marginal distributions, (ii) all the
72 72 72 72
9 11 13 33
conditional distributions and (iii) the probability distribution of (𝑋 + 𝑌). 3
72 72 72 72
Solution:
Marginal 18 24 30
To find the value of k 1
Distribution of X 72 72 72

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 3
 ∑ 𝑝(𝑥, 𝑦) (𝑝( 𝑋) 𝑋 + 𝑌 Probability
𝑦 = 𝑥) 3
1 𝑃(0,1) =
(ii)The conditional distribution of X given 𝑌 = 𝑦 is 72
𝑃[𝑋 = 𝑥 ∩ 𝑌 = 𝑦] 5 6 11
𝑃[𝑋 = 𝑥/𝑌 = 𝑦] =
2 𝑃(1,1) + 𝑃(0,2) = + =
𝑃[𝑌 = 𝑦] 72 72 72
7 8 9
X 𝑃(2,1) + 𝑃(1,2) + 𝑃(0,3) = + +
0 1 2 72 72 72
Y 3⁄ 3
72 = 3 24
(0,1) = =
3 5 7 15⁄ 15 72
1 72
15 15 15 11 10 21
4 𝑃(1,3) + 𝑃(2,2) = + =
6 8 10 72 72 72
2
24 24 24 13
5 𝑃(2,3) =
9 11 13 72
3
33 33 33 Problem: 4
Given the following bivariate probability distribution obtain
The conditional distribution of Y given 𝑋 = 𝑥 is (i) Marginal distributions of X and Y
𝑃[𝑋 = 𝑥 ∩ 𝑌 = 𝑦]
𝑃[𝑌 = 𝑦/𝑋 = 𝑥] = (ii) Conditional distribution of X given Y=2 and
𝑃[𝑋 = 𝑥]
(iii) The conditional distribution of Y given X=1.
X
0 1 2 X -1 0 1
Y
Y
3⁄
3 5 7 72 = 3
1 (0,1) =
18⁄ 18 1 2 1
18 24 30 72 0
15 15 15
6 8 10
2
18 24 30 1 3 2 1
9 11 13 15 15 15
3
18 24 30 2 2 1 2
15 15 15
(iii) Probability distribution of X+Y
Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 4
 𝑃[𝑋 = 𝑥 ∩ 𝑌 = 2]
𝑃[𝑋 = 𝑥/𝑌 = 2] = 𝑤ℎ𝑒𝑟𝑒 𝑥 = −1,0,1
Solution: 𝑃[𝑌 = 2]
2/15 2
𝑃[𝑋 = −1/𝑌 = 2] = =
X -1 0 1 ∑ 𝑝(𝑥, 𝑦) 5/15 5
𝑥 1/15 1 2/15 2
𝑃[𝑋 = 0/𝑌 = 2] = 5/15 = 5and𝑃[𝑋 = 1/𝑌 = 2] = 5/15 = 5
Y
𝑋: -1 0 1
0 1 2 1 4
15 15 15 15 𝑃 (𝑋/𝑌 2 1 2
= 2): 5 5 5
1 3 2 1 6
15 15 15 15
2 2 1 2 5 (iii) Conditional Distribution of Y given X=1:
15 15 15 15
𝑃[𝑌 = 𝑦 ∩ 𝑋 = 1]
6 5 4 1 𝑃[𝑌 = 𝑦/𝑋 = 1] = 𝑤ℎ𝑒𝑟𝑒 𝑦 = 0,1,2
∑ 𝑝(𝑥, 𝑦) 𝑃[𝑋 = 1]
𝑦 15 15 15
𝑋: 0 1 2
(i) Marginal Distribution of X and Y: 𝑃 (𝑌/𝑋 1 1 2
𝑃[𝑋 = 𝑥] = ∑ 𝑝(𝑥, 𝑦) 𝑎𝑛𝑑 𝑃[𝑌 = 𝑦] = ∑ 𝑝(𝑥, 𝑦) = 1): 4 4 4
𝑦 𝑦

𝑥: -1 0 1
𝑝( 𝑥 ) : 6 5 4
Problems Based on Continuous Random Variables
15 15 15
Problem: 1 Show that the function𝑓(𝑥, 𝑦) =
2
(2𝑥 + 3𝑦),0 < 𝑥 < 1, 0 < 𝑦 < 1
𝑦: 0 1 2 {5
0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑝 (𝑦 ): 4 6 5
is a joint density functions of x and y.
15 15 15
Solution:
We know that 𝑓(𝑥, 𝑦) satisfies the conditions
(ii) Conditional Distribution of X given 𝑌 = 2:
(i) 𝑓(𝑥, 𝑦) ≥ 0

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 5
 ∞ ∞ 5
(ii) (ii) ∫−∞ ∫−∞ 𝑓 (𝑥, 𝑦)𝑑𝑥𝑑𝑦 = 1, 𝑡ℎ𝑒𝑛 𝑓(𝑥, 𝑦) 𝑖𝑠 𝑎 𝑗. 𝑑. 𝑓. 5
𝑦2
−4𝑘 ∫ (1 − 𝑦)𝑑𝑦 = 1  − 4𝑘 [𝑦 − ] = 1
Now, 1 2 1

(i) 𝑓(𝑥, 𝑦) ≥ 0 in the interval 0 < 𝑥 < 1, 0 < 𝑦 < 1 15 1

−4𝑘 [(− ) − ( )] = 1 − 4𝑘 (−8) = 1
2 2
1
∞ ∞ 1 12
(ii) ∫−∞ ∫−∞ 𝑓 (𝑥, 𝑦)𝑑𝑥𝑑𝑦 = ∫0 ∫0 5 (2𝑥 + 3𝑦) 𝑑𝑥𝑑𝑦 𝐾= ( )
32
2 1 𝑥2
1 Problem:3 The joint p.d.f of R.V 𝑋 𝑎𝑛𝑑 𝑌 is given by f(𝑥, 𝑦) =
=5 ∫0 [2 2 + 3𝑦𝑥] dy 2 +𝑦 2 )
0
𝑘𝑥𝑦𝑒 −(𝑥 , 𝑥 > 0, 𝑦 > 0. Find the value of 𝑘.Prove that 𝑋 𝑎𝑛𝑑 𝑌 are
1
2 1 2 3𝑦 2
=5 ∫0 (1 + 3𝑦) 𝑑𝑦 = 5 [𝑦 + 2
] independent.
0

2 3 2 5 Solution:
= (1 + ) = [ ] = 1 ∞ ∞
5 2 5 2 Since 𝑓(𝑥, 𝑦) is a p.d.f, we have∫−∞ ∫−∞ 𝑓 (𝑥, 𝑦) 𝑑𝑥𝑑𝑦 = 1
∞ ∞
∞ ∞
∫ ∫ 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 = 1 ∫ ∫ 𝑘𝑥𝑦𝑒 −(𝑥
2 +𝑦 2 )
−∞ −∞
𝑑𝑥𝑑𝑦 = 1
0 0
Since 𝑓(𝑥, 𝑦) satisfies the given two conditions, it is a joint distribution ∞ ∞
2 2
𝑘 ∫ ∫ 𝑥𝑦𝑒 −𝑥 . 𝑒 −𝑦 𝑑𝑥𝑑𝑦 = 1
function. 0 0

Problem: 2 Find ’k’ if the joint probability density function of a bivariate ∞ ∞

2 2
𝑘 ∫ 𝑦𝑒 −𝑦 𝑑𝑦 ∫ 𝑥𝑒 −𝑥 𝑑𝑥 = 1
random variable (x, y) is given by 𝑓 (𝑥, 𝑦) = 0 0
2
{
𝑘(1 − 𝑥)(1 − 𝑦), 𝑖𝑓0 < 𝑥 < 4; 1 < 𝑦 < 5 𝑃𝑢𝑡 𝑥 = 𝑡 𝑤ℎ𝑒𝑛 𝑥 = 0 𝑃𝑢𝑡 𝑦 2 = 𝑠 𝑤ℎ𝑒𝑛 𝑦 = 0
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑡ℎ𝑒𝑛 𝑡 = 0 𝑡ℎ𝑒𝑛 𝑠 = 0
Solution:
∞ ∞ 2𝑥𝑑𝑥 = 𝑑𝑡 𝑤ℎ𝑒𝑛 𝑥 = ∞ 2𝑦𝑑𝑦 = 𝑑𝑠 𝑤ℎ𝑒𝑛 𝑦 = ∞
Since 𝑓(𝑥, 𝑦) is a p.d.f, we have ∫−∞ ∫−∞ 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 = 1
𝑡ℎ𝑒𝑛 𝑡 = ∞ 𝑡ℎ𝑒𝑛 𝑠 = ∞
5 4
∫ ∫ 𝑘 (1 − 𝑥)(1 − 𝑦)𝑑𝑥𝑑𝑦 = 1 1 1
1 0 ⟹ 𝑥𝑑𝑥 = 𝑑𝑡 ⟹ 𝑦𝑑𝑦 = 𝑑𝑠
2 2
5 4
∫ {𝑘(1 − 𝑦) ∫ (1 − 𝑥)𝑑𝑥} 𝑑𝑦 = 1 The above equation becomes
1 0
1 ∞ −𝑡 1 ∞ −𝑠
5
𝑥2
4 𝑘 { ∫ 𝑒 𝑑𝑡} { ∫ 𝑒 𝑑𝑠} = 1
∫ {𝑘(1 − 𝑦) [𝑥 − ]} 𝑑𝑦 = 1 2 0 2 0
1 2 0

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 6
 ∞ ∞ 8𝑥𝑦 2𝑦
𝑘 𝑒 −𝑡 𝑒 −𝑠 𝑘 𝑘 = = 𝑥 2 0 < 𝑦 < 𝑥 , 0 < 𝑥 < 1.
⟹ [ ] [ ] = 1 ⟹ [0 − 1][0 − 1] = 1 ⟹ = 1 ∴ 𝑘 = 4 4𝑥 3
4 −1 0 −1 0 4 4

To prove that 𝑋 𝑎𝑛𝑑 𝑌 are independent, We have to prove 𝐶𝑥 (𝑥 − 𝑦), 0 < 𝑥 < 2, −𝑥 < 𝑦 < 𝑥
Problem: 5 Given𝑓(𝑥, 𝑦) = { Find (i) C,
𝑓𝑋 (𝑥). 𝑓𝑌 (𝑦) = 𝑓(𝑥, 𝑦) 0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
∞
By formula, 𝑓𝑋 (𝑥) = ∫−∞ 𝑓 (𝑥, 𝑦)𝑑𝑦 (ii) 𝑓(𝑥), (iii) 𝑓(𝑦/𝑥)
∞ 2 +𝑦 2 ) Solution:
= ∫0 𝑘𝑥𝑦𝑒 −(𝑥 𝑑𝑦 ∵ 𝑦 > 0 𝑖. 𝑒. , 𝑦 𝑣𝑎𝑟𝑖𝑒𝑠 0 𝑡𝑜 ∞
∞
(i) We know that the function 𝑓 (𝑥, 𝑦) should satisfy
−𝑥 2 −𝑦 2 −𝑥 2
1 ∞ ∞
= 4𝑥𝑒 ∫ 𝑦𝑒 𝑑𝑦 = 4𝑥𝑒 ( ) ∫−∞ ∫−∞ 𝑓(𝑥, 𝑦) 𝑑𝑦 𝑑𝑥 = 1
0 2
2 𝑥
−𝑥 2 Now ∫0 ∫−𝑥 𝐶𝑥 (𝑥 − 𝑦) 𝑑𝑦 𝑑𝑥 = 1
𝑓𝑋 (𝑥) = 2𝑥𝑒 ,𝑥 > 0
2 2 𝑥 2
Similarly, we get 𝑓𝑌 (𝑦) = 2𝑦𝑒 −𝑦 , 𝑦 > 0 𝑦2
𝑥
⟹ 𝐶 ∫ ∫ 𝑥(𝑥 − 𝑦) 𝑑𝑦 𝑑𝑥 = 1 ⟹ 𝐶 ∫ 𝑥 [(𝑥𝑦 − ) ] 𝑑𝑥 = 1
2
Now𝑓𝑋 (𝑥). 𝑓𝑌 (𝑦) = 2𝑥𝑒 −𝑥 . 2𝑦𝑒 −𝑦 = 4𝑥𝑦𝑒 −(𝑥
2 2 +𝑦 2 )
, 𝑥 > 0, 𝑦 > 0 2 −𝑥
0 −𝑥 0
= 𝑓(𝑥, 𝑦) ∴ 𝑋 𝑎𝑛𝑑 𝑌 𝑎𝑟𝑒 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡. 2
𝑥2 𝑥2
Problem: 4 The joint p.d.f of the random variables X and Y is given by ⟹ 𝐶 ∫ 𝑥 [(𝑥 2 − ) − (−𝑥 2 − )] 𝑑𝑥 = 1
2 2
0
8𝑥𝑦, 0 < 𝑥 < 1, 0 < 𝑦 < 𝑥
𝑓 (𝑥, 𝑦) = { Find(𝑖)𝑓𝑋 (𝑥) (𝑖𝑖) 𝑓𝑌 (𝑦) (𝑖𝑖𝑖) 𝑓(𝑦/𝑥) 2 2
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
⟹ 𝐶 ∫ 𝑥[2𝑥 2 ] 𝑑𝑥 = 1 ⇒ 2𝐶 ∫ 𝑥 3 𝑑𝑥 = 1
Solution:
0 0
∞
The m.d.fof X is 𝑓𝑋 (𝑥) = 𝑓 (𝑥) = ∫−∞ 𝑓(𝑥, 𝑦) 𝑑𝑦 2𝐶 4 2 𝐶 1
⇒ [𝑥 ]0 = 1 ⇒ (16) = 1 ⇒ 8𝐶 = 1 ∴ 𝐶 =
𝑓 (𝑥) =
𝑥 4 2 8
∫0 8𝑥𝑦 𝑑𝑦
1
x
= 8x ∫0 y dy =
8x
[y 2 ]x0 = 4x(x 2 ) = 4x 3 𝑓(𝑥, 𝑦) = 𝑥(𝑥 − 𝑦), 0 < 𝑥 < 2, −𝑥 < 𝑦 < 𝑥
2 8
X ∞
=
1
∫0 8𝑥𝑦 𝑑𝑥 , 0<𝑥<1 = X (ii) m.d.f of ‘X’ is 𝑓 (𝑥) = 𝑓𝑋 (𝑥) = ∫−∞ 𝑓 (𝑥, 𝑦)𝑑𝑦
Y = 𝑥 𝑥
1
1 1 2 1 2 𝑦2
8𝑦 2 1 ( )
= ∫ 𝑥 − 𝑥𝑦 𝑑𝑦 = [𝑥 𝑦 − 𝑥 ]
= 8𝑦 ∫ 𝑥𝑑𝑥 = [𝑥 ]𝑦 = 4𝑦(1 − 𝑦 2 ) Y= 8 8 2 −𝑥
2 −𝑥
𝑦 0
𝑓(𝑥,𝑦) 1 𝑥3 𝑥3
WKT by conditional probability distribution is 𝑓(𝑦/𝑥) = = [ (𝑥 3 − ) − (−𝑥 3 − )
𝑓𝑋 (𝑥) 8 2 2

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 7
 1 𝑥3 3
= (2𝑥 3 ) = 𝑃(𝑋 < 1 ∩ 𝑌 < 3) = − − − − − −(1)
8 4 8
1
𝑓(𝑥,𝑦) 𝑥(𝑥−𝑦) 1 4 1
8
(iii)𝑓(𝑦/𝑥) = = 𝑥3
= 𝑥 (𝑥 − 𝑦). = (𝑥 − 𝑦), −𝑥 < 𝑦 < 𝑥
𝑓𝑋 (𝑥) 8 𝑥3 2𝑥 2
4 (ii) To evaluate 𝑃(𝑋 < 1/𝑌 < 3)
Problem: 6 If the joint p.d.f of X and Y is given 𝑓(𝑥,𝑦)
W.K.T by condionally density function of x given y is 𝑓(𝑥/𝑦) = fY (y)
1
( )
𝑓(𝑥, 𝑦) = {8 6 − 𝑥 − 𝑦 , 0 < 𝑥 < 2, 2 < 𝑦 < 4
𝑃(𝑋<1,𝑌<3)
Now, 𝑃(𝑋 < 1/𝑌 < 3) = − − − −(2)
𝑝(𝑌<3)
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
3 3 2
Find (𝑖) 𝑃(𝑋 < 1 ∩ 𝑌 < 3), (𝑖𝑖 ) 𝑃(𝑋 < 1/𝑌 < 3), (𝑖𝑖𝑖) 𝑃(𝑋 + 𝑌 < 3) 1
𝑃 (𝑌 < 3) = ∫ fY (y)dy = ∫ ∫ (6 − 𝑥 − 𝑦) 𝑑𝑥 𝑑𝑦
8
Solution: 2 2 0
2 3
(i) To evaluate 𝑃(𝑋 < 1 ∩ 𝑌 < 3) 1
𝑏 𝑏 = ∫[∫(6 − 𝑥 − 𝑦) 𝑑𝑦] 𝑑𝑥
We know that 𝑃[(𝑎1 ≤ 𝑋 ≤ 𝑏1 )⋂(𝑎2 ≤ 𝑌 ≤ 𝑏2 )] = ∫𝑎 1 ∫𝑎 2 𝑓(𝑥, 𝑦) 𝑑𝑦 𝑑𝑥 8
1 2 0 2
1 3 2
1 1 𝑦2
𝑃(𝑋 < 1 ∩ 𝑌 < 3) = ∫ ∫ (6 − 𝑥 − 𝑦) 𝑑𝑦𝑑𝑥 = ∫[(6 − 𝑥)𝑦 − ]³₂ 𝑑𝑥
8 8 2
0 2 0
3 1 2
3
1 1 1 𝑦2 1 9
= ∫ ∫(6 − 𝑥 − 𝑦) 𝑑𝑦 𝑑𝑥 = ∫ [(6 − 𝑥)𝑦 − ] 𝑑𝑥 = ∫[(3(6 − 𝑥) − ) − (2(6 − 𝑥) − 2)] 𝑑𝑥
8 0 8 2 8 2
2 0 2 0
1 2
1 9 1 9
= ∫[(3(6 − 𝑥) − ) − (2(6 − 𝑥) − 2)] 𝑑𝑥 = ∫[18 − 3𝑥 − − 12 + 2𝑥 + 2] 𝑑𝑥
8 2 8 2
0 0
1 2 2
1 9 1 9 1 7
= ∫[18 − 3𝑥 − − 12 + 2𝑥 + 2] 𝑑𝑥 = ∫[8 − − 𝑥] 𝑑𝑥 = ∫( − 𝑥) 𝑑𝑥
8 2 8 2 8 2
0 0 0
1 1 1 7 𝑥2 1
1 9 1 7 = [ 𝑥 − ]20 = (7 − 2)
= ∫[8 − − 𝑥] 𝑑𝑥 = ∫( − 𝑥) 𝑑𝑥 8 2 2 8
8 2 8 2
0 0 5
2 𝑃(𝑌 < 3) = − − − − − − (3)
1 7 𝑥 1 7 1 1 6 8
= [ 𝑥 − ]10 = [ − ] = ( ) 3/8
8 2 2 8 2 2 8 2 𝑃(𝑋 < 1/𝑌 < 3) = = 3/5[ from(1) &(2)]
5/8

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 8
 3 3−𝑦 1
(ii) 𝑃(𝑋 + 𝑌 < 3) = ∫2 ∫0 (6 − 𝑥 − 𝑦) 𝑑𝑥 𝑑𝑦 (iii)Verify whether 𝑋 𝑎𝑛𝑑 𝑌 are independent.
8
1 3 3−𝑦 (iv)Find 𝐸(𝑋) 𝑎𝑛𝑑 𝐸(𝑌)
= ∫ [∫ (6 −
8 2 0
𝑥 − 𝑦) 𝑑𝑥] 𝑑𝑦
Solution: Y
3
2 3
1 𝑥 3−𝑦 Given that𝑓(𝑥, 𝑦) = 8𝑥𝑦, 0 < 𝑥 < 𝑦 < 1 =
= ∫[(6 − 𝑦)𝑥 − ]0 𝑑𝑦 (0, - X+y 1
8 2 3) =3 (i)The marginal density function of ‘𝑥’ is
2 x X
3 (1, ∞ X
𝑓𝑋 (𝑥) = 𝑓(𝑥) = ∫−∞ 𝑓(𝑥, 𝑦)𝑑𝑦 (𝑓𝑜𝑟𝑚𝑢𝑙𝑎 ) =
1 (3 − 𝑦)² 2) =
= ∫[(6 − 𝑦)(3 − 𝑦) − ] 𝑑𝑦 2 0
8 2 1 𝑦2
1
Y
2 𝑓 (𝑥 ) = ∫𝑥 8𝑥𝑦𝑑𝑦 = 8𝑥 ( 2 ) = 4𝑥(1 − 𝑥 2 )
𝑥
3 (3,0)
1 1 2
4𝑥(1 − 𝑥 ),0 < 𝑥 < 1
= ∫[18 − 9𝑦 + 𝑦 2 − (3 − 𝑦)² 𝑑𝑦 𝑓𝑋 (𝑥) = 𝑓 (𝑥) = {
8 2 0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
2

1 9 𝑦3 1 The marginal density function of ‘𝑦’ is

= [18𝑦 − 𝑦 2 + − (3 − 𝑦)³]32 ∞
8 2 3 −2 × 3 𝑓𝑌 (𝑦) = 𝑓(𝑦) = ∫−∞ 𝑓 (𝑥, 𝑦)𝑑𝑥 (𝑓𝑜𝑟𝑚𝑢𝑙𝑎) Y
1 81 1 8 1 =
= [(54 − + 9 + (0)) − (36 − 18 + + )] X=0 1
8 2 6 3 6 X=Y
𝑦 2 𝑦
1 81 17 𝑥
= [(63 − ) − (18 + )] = ∫ 8𝑥𝑦 𝑑𝑥 = 8𝑦 ( ) = 4𝑦 3
8 2 6 0 2 0
1 126 − 81 108 + 17 (0,
3
4𝑦 , 0 < 𝑦 < 1
= [(
8 2
)−(
6
)] 𝑓𝑌 (𝑦) = 𝑓(𝑦) = { − − − − − − − (1) 0)
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
1 45 125 1 1 3
= [ − ] (ii)To find𝑃(𝑋 < / < 𝑌 < )
8 2 6 4 2 4
𝑓(𝑥,𝑦)
1 135 − 125 1 10 The conditionally density function of X given Y is 𝑓(𝑥/𝑦) = 𝑓
= [ ] = ( ) = 5/24 𝑌 (𝑦)
8 6 8 6
1 1 3
1 1 3 𝑃(𝑋< ∩ <𝑌< )
4 2 4
Now,𝑃 (𝑋 < / < 𝑌 < ) = 1 3 − − − − − − − − − (2)
4 2 4 𝑃( <𝑌< )
2 4
Problem: 7Let X and Y be two random variable s with the joint
1 1 3
8𝑥𝑦, 0 < 𝑥 < 𝑦 < 1 First we find 𝑃 (𝑋 < ∩ < 𝑌 < )
4 2 4
p.d.f 𝑓 (𝑥, 𝑦) = {
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
(i) Find the marginal probability density function of X and Y.
1 1 3
(ii) Find 𝑃(𝑋 < 4 / 2 < 𝑌 < 4)

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 9
 ∞
Using formula,𝐸 (𝑋) = ∫−∞ 𝑥𝑓 (𝑥)𝑑𝑥
𝑏 𝑏
We know that 𝑃[(𝑎1 ≤ 𝑋 ≤ 𝑏1 )⋂(𝑎2 ≤ 𝑌 ≤ 𝑏2 )] = ∫𝑎 1 ∫𝑎 2 𝑓(𝑥, 𝑦) 𝑑𝑦 𝑑𝑥 1
1 2
= ∫ 𝑥 × 4 × 𝑥 (1 − 𝑥 2 )𝑑𝑥 ∵ 0 < 𝑥 < 1
1/4 3/4 0
1 1 3
𝑃 (𝑋 < ∩ < 𝑌 < ) = ∫ ∫ 8𝑥𝑦𝑑𝑦𝑑𝑥 1 1
4 2 4 0 1/2 2 4
𝑥3 𝑥5 2 8
( )
= 4 ∫ 𝑥 − 𝑥 𝑑𝑥 = 4 ( − ) = 4 ( ) =
1/4 3/4 1/4 0 3 5 0 15 15
𝑥 2 3/4
= 8∫ 𝑥 (∫ 𝑦𝑑𝑦) 𝑑𝑥 = 8 ∫ (𝑦 )1/2 𝑑𝑥
0 1/2 0 2 To find 𝐸(𝑌):
1 1 ∞
4 9 1 4 5 Using formula 𝐸(𝑌) = ∫−∞ 𝑦𝑓(𝑦)𝑑𝑦
= 4 ∫ 𝑥 ( − ) 𝑑𝑥 = 4 ∫ 𝑥 ( ) 𝑑𝑥 1
0 16 4 0 16 1 1
𝑦5 4
1/4
= ∫ 𝑦4𝑦 3 𝑑𝑦 = 4 ∫ 𝑦 4 𝑑𝑦 = 4 ( ) =
5 𝑥2 5 1 5 0 0 5 0 5
= 4× ( ) = ( ) = − − − − − −(3)
16 2 0 8 16 128 Problem: 8 Given the joint density function 𝑓 (𝑥, 𝑦) =
1 3 3/4 (1+3𝑦 2 )
𝑃 ( < 𝑌 < ) = ∫ 𝑓 (𝑦)𝑑𝑦 {𝑥 4
, 0 < 𝑥 < 2, 0 < 𝑦 < 1. Find the marginal densities 𝑔(𝑥), ℎ(𝑦) and
2 4 1/2 0, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
3/4 4 3/4 1 1 1
𝑦 the conditional density 𝑓(𝑥 ⁄ 𝑦) and evaluate 𝑃[ < 𝑥 < ⁄𝑦 = ]
=∫ 4𝑦 3 𝑑𝑦 = 4 ( ) 4 2 3
1/2 4 1/2
Solution:
4 81 1 65 The marginal density function of ‘𝑥’ is
= ( − )= − − − − − − − − − − − (4)
4 256 16 256 ∞
Sub.(3) and (4) in (2) we get 𝑔(𝑥) = ∫ 𝑓(𝑥, 𝑦) 𝑑𝑦
5 −∞
1 1 3 128 2
𝑃 (𝑋 < / < 𝑌 < ) = 65 = 1
𝑥 (1 + 3𝑦 2 ) 𝑥 𝑥 𝑥
4 2 4 13 = ∫ 𝑑𝑦 = [𝑦 + 𝑦 3 ]10 = [1 + 1] =
256 4 4 4 2
0
(iii)To check X and Y are independent 𝑥
𝑔(𝑥) = ,0 < 𝑥 < 2
Using formula,𝑓𝑋 (𝑥). 𝑓𝑌 (𝑦) = 𝑓(𝑥, 𝑦) 2
𝑓𝑋 (𝑥). 𝑓𝑌 (𝑦) = 4𝑥 (1 − 𝑥 2 ). 4𝑦 3 ≠ 𝑓 (𝑥, 𝑦) The marginal density function of ‘𝑦’ is
∞
∴ 𝑋 𝑎𝑛𝑑 𝑌are not independent.
ℎ(𝑦) = ∫ 𝑓(𝑥, 𝑦) 𝑑𝑥
(iv)To find 𝐸 (𝑋): −∞

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 10
 2
2
𝑥 (1 + 3𝑦 2 ) 1 + 3𝑦 2 𝑥 2 1 + 3𝑦 2 Given 𝐹 (𝑥, 𝑦) = (1 − 𝑒 −𝑥 )(1 − 𝑒 −𝑦 ) = 1 − 𝑒 −𝑥 − 𝑒 −𝑦 + 𝑒 −𝑥 𝑒 −𝑦
= ∫ 𝑑𝑥 = [ ] =
0 4 4 2 0 2 𝜕2 𝐹(𝑥,𝑦)
The joint p.d.f of 𝐹(𝑥, 𝑦) is 𝑓(𝑥, 𝑦) = 𝜕𝑥𝜕𝑦
1 + 3𝑦 2
ℎ(𝑦) = , 0<𝑦<1 𝜕 2 [1 − 𝑒 −𝑥 − 𝑒 −𝑦 + 𝑒 −𝑥 𝑒 −𝑦 ]
2 =
𝜕𝑥𝜕𝑦
The conditional density function of 𝑋 given 𝑌 = 𝑦 is
𝜕 −𝑦
𝑓(𝑥 , 𝑦) 𝑥 (1 + 3𝑦 2 ) 2 𝑥 = [𝑒 − 𝑒 −𝑥 𝑒 −𝑦 ] = 𝑒 −𝑥 𝑒 −𝑦
𝑓(𝑥⁄𝑦) = = = 𝜕𝑥
ℎ(𝑦) 4 1 + 3𝑦 2 2
= 𝑒 −(𝑥+𝑦)
𝑥 1+3𝑦 2
Since 𝑔(𝑥) ℎ(𝑦) = 2 2
= 𝑓(𝑥 , 𝑦), X and Y are independent 𝑒 −(𝑥+𝑦) 𝑓𝑜𝑟𝑥 > 0, 𝑦 > 0
𝑓 (𝑥, 𝑦) = {
𝑃 (𝐴 ∩ 𝐵) 𝑃 (𝐴) 𝑃 (𝐵) 0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑃(𝐴⁄𝐵) = = = 𝑃(𝐴)
𝑃 (𝐵) 𝑃 (𝐵) (i) The m.d.f. of x is
∞
Therefore
𝑓𝑥 (𝑥) = 𝑓(𝑥) = ∫ 𝑓(𝑥, 𝑦)𝑑𝑦
1 1 −∞
<𝑋< 1 1
𝑝 (4 2⁄
1) = 𝑃 ( < 𝑋 < )
∞ ∞
𝑌= 2 4 2 = ∫ 𝑒 −(𝑥+𝑦) 𝑑𝑦 = ∫ 𝑒 −𝑥 𝑒 −𝑦 𝑑𝑦(0 < 𝑦 < ∞)
0 0
1/2 1/2 2 1/2
𝑥 𝑥 1 1 1 3 ∞
∞
= ∫ 𝑔(𝑥) 𝑑𝑥 = ∫ 𝑑𝑥 = [ ] = [ − ]= = 𝑒 −𝑥 ∫ 𝑒 −𝑦 𝑑𝑦 = −𝑒 −𝑥 [𝑒 −𝑦 ] = −𝑒 −𝑥 [0 − 1]
1/4 1/4 2 4 1/4 4 4 16 64 0 0
−𝑥
=𝑒 , 𝑥 > 0 ∵ [ 𝑒 −∞ = 0&𝑒 0 = 1]
REMARK: The m.d.f of y is
∞ ∞
Let 𝐹𝑥𝑦 (𝑥, 𝑦) be the joint p.d.f. Then the joint p.d.f of X and Y is given by,
𝑓𝑦 (𝑥) = 𝑓(𝑦) = ∫ 𝑓(𝑥, 𝑦)𝑑𝑥 = ∫ 𝑒 −(𝑥+𝑦) 𝑑𝑥
𝜕 2 (𝐹(𝑥, 𝑦)) −∞ 0
𝑓(𝑥, 𝑦) = ∞
𝜕𝑥𝜕𝑦 = ∫ 𝑒 −𝑥 𝑒 −𝑦 𝑑𝑥
0
Problem: 9 If the joint distribution function of X and Y is given by
∞
𝑒 −𝑥 ∞
(1 − 𝑒 −𝑥 )(1 − 𝑒 −𝑦 ) 𝑓𝑜𝑟 𝑥 > 0, 𝑦 > 0 = 𝑒 −𝑦 ∫ 𝑒 −𝑥 𝑑𝑥 = 𝑒 −𝑦 [ ] = −𝑒 −𝑦 [0 − 1]
𝐹(𝑥, 𝑦) = { −1 0
0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 0
−𝑦
(i) Find the marginal densities of X and Y. =𝑒 ,𝑦 > 0

(ii )Are X and Y independent?. (iii) Find 𝑃(1 < 𝑋 < 3 , 1 < 𝑌 < 2)
Solution: (ii) Next, to check X and Y are independent,

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 11
Now, 𝑓𝑋 (𝑥). 𝑓𝑌 (𝑦) = 𝑒 −𝑥 . 𝑒 −𝑦 = 𝑒 −(𝑥+𝑦) = 𝑓(𝑥, 𝑦) (ii)Conditional Density Functions:
X and Y are independent. 𝑥 𝑓(𝑥, 𝑦) 2 1
𝑓( ) = = = ,0 < 𝑥 < 𝑦
𝑦 𝑓𝑌 (𝑦) 2𝑦 𝑦
(iii) To evaluate 𝑃(1 < 𝑋 < 3 , 1 < 𝑌 < 2)
3 2 𝑦 𝑓 (𝑥, 𝑦) 2 1
−𝑥 −𝑦 𝑓( ) = = = ,0 ≤ 𝑥 < 𝑦 ≤ 1
𝑃 (1 < 𝑋 < 3 , 1 < 𝑌 < 2) = ∫ ∫ 𝑒 𝑒 𝑑𝑦𝑑𝑥 𝑥 𝑓𝑋 (𝑥) 2(1 − 𝑥) 1 − 𝑥
1 1
3 2 3 2
(iii)Conditional mean of X given 𝑌 = 𝑦
−(𝑥+𝑦) −𝑥 −𝑦
=∫ ∫ 𝑒 𝑑𝑦𝑑𝑥 = ∫ 𝑒 [∫ 𝑒 𝑑𝑦] 𝑑𝑥 𝑥 ∞
1 1 1 1 𝐸[𝑥/𝑦] = ∫ 𝑥𝑓 ( ) 𝑑𝑥
−∞ 𝑦
3 3
= − ∫ 𝑒 −𝑥 [𝑒 −2 − 𝑒 −1 ]𝑑𝑥 = (𝑒 −1 − 𝑒 −2) ∫ 𝑒 −𝑥 𝑑𝑥 𝑦
1 1 𝑥2 𝑦 𝑦
1 1 = ∫ 𝑥 ( ) 𝑑𝑥 = [ ]0 =
0 𝑦 𝑦 2 2
−1 −2 )( −3 −1 ) −4 −2
= (𝑒 −𝑒 𝑒 −𝑒 = 𝑒 −𝑒 − 𝑒 −5 + 𝑒 −3 . 𝑦
𝑦
2
𝑥 1 𝑥3
2
𝑦2
𝐸[𝑥 /𝑦 ] = ∫ 𝑥 𝑓 ( ) 𝑑𝑥 = [ ] = .
0 𝑦 𝑦 3 0 3
Problem 10:
Conditional variance 𝑉[𝑥/𝑦 ] = 𝐸[𝑥 2 /𝑦] − {𝐸[𝑥/𝑦]}2
If the joint p.d.f of (X,Y) is given by 𝑓 (𝑥, 𝑦) = 2, 0 ≤ 𝑥 < 𝑦 ≤ 1, find
𝑦2 𝑦2 𝑦2
(i)Marginal density functions of X and Y. = − =
3 4 12
𝑥 𝑦
(ii)Conditional densities 𝑓 (𝑦) and 𝑓 (𝑥 ). 𝑦2 1
Conditional Variance of X given 𝑌 = 2 = 12 | = 3.
𝑦=2
(iii)Conditional Variance of X given 𝑌 = 2.
Problem 11:
Solution:
The joint p.d.f of two random variables X and Y is 𝑓 (𝑥, 𝑦) =
(i)Marginal density functions 𝑓𝑋 (𝑥)and𝑓𝑌 (𝑦):
𝑘𝑥 (𝑥 − 𝑦), 0 < 𝑥 < 2, |𝑦| < 𝑥
∞ {
0, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
𝑓𝑋 (𝑥) = ∫ 𝑓 (𝑥, 𝑦)𝑑𝑦
−∞ Evaluate the constant k. Find the marginal and conditional density of Y
1
= ∫ 2𝑑𝑦 = 2(1 − 𝑥), 0 ≤ 𝑥 ≤ 1 given X.
𝑥
Solution:
∞
𝑓𝑌 (𝑦) = ∫ 𝑓(𝑥, 𝑦)𝑑𝑥 |𝑦| < 𝑥 ⇒ −𝑥 < 𝑦 < 𝑥.
−∞
𝑦 (i) To find k
= ∫ 2𝑑𝑥 = 2𝑦, 0 ≤ 𝑦 ≤ 1 2 𝑥
0 ∫ ∫ 𝑘(𝑥 2 − 𝑥𝑦)𝑑𝑦𝑑𝑥 = 1
0 −𝑥
Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 12
 2 2
1 𝑦 5𝑦 3
⇒ 𝑘 ∫ 𝑥 2 (2𝑥)𝑑𝑥 − 𝑘 ∫ 𝑥 (0)𝑑𝑥 = 1 = − + 𝑖𝑛 − 2 < 𝑦 < 0
0 0 3 4 48
2 2
𝑥 4 2 1 𝑥3 𝑥2
∴ 2𝑘 [ ] = 1 and ℎ(𝑦) = [[ ] − 𝑦 [ ] ]
4 0 8 3 𝑦 2 𝑦

1 1 8 𝑦3 𝑦3
∴ 8𝑘 = 1 ⇒ 𝑘 = = [ − − 2𝑦 + ]
8 8 3 3 2
(ii) Marginal densities of X and Y: 1 𝑦 𝑦3
∞ = − + 𝑖𝑛 0 < 𝑦 < 2
𝑔(𝑥) = ∫−∞ 𝑓(𝑥, 𝑦)𝑑𝑦 3 4 48

1 𝑥 ∴ The Marginal density of Y is

= ∫ (𝑥 2 − 𝑥𝑦)𝑑𝑦 (2,2)
8 −𝑥 1 𝑦 5𝑦 3
− + 𝑖𝑛 − 2 < 𝑦 < 0
𝑥2 𝑥 𝑥 𝑥 y=x ℎ(𝑦) = 3 4 48 3
= ∫ 𝑑𝑦 − ∫ 𝑦𝑑𝑦 X=2 1 𝑦 𝑦
8 −𝑥 8 −𝑥 { 3 − 4 + 48 𝑖𝑛 0 < 𝑦 < 2
(2,0)
𝑥2 𝑥
𝑥 𝑥3 Conditional density function
= [𝑦]−𝑥 − (0) = .
8 8 4
Y=-x 𝑓 (𝑥, 𝑦)
∴ The Marginal density of X is 𝑓(𝑦/𝑥) =
𝑔(𝑥)
𝑥3 (2,- 𝑥
𝑔(𝑥) = ,0 < 𝑥 < 2 2) 𝑦 (𝑥 − 𝑦) 𝑥 − 𝑦
4 ∴ 𝑓 ( ) = 8 𝑥3 = , −𝑥 < 𝑦 < 𝑥, 0 < 𝑥 < 2
∞ 𝑥 2𝑥 2
4
ℎ(𝑦) = ∫ 𝑓(𝑥, 𝑦)𝑑𝑥
−∞ Problem 12:
2
1 −(𝑥2 +𝑦2 )
∴ ℎ(𝑦0 = ∫ (𝑥 2 − 𝑥𝑦)𝑑𝑥(−2 < 𝑦 < 0) The joint p.d.f of two random variables X and Y is 𝑓 (𝑥, 𝑦) = { 𝑥𝑦𝑒 2 𝑑𝑦 ,
8 −𝑦
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
1 2 𝑥 ≥ 0, 𝑦 ≥ 0
𝑎𝑛𝑑 ℎ(𝑦) = ∫ (𝑥 2 − 𝑥𝑦)𝑑𝑥(0 < 𝑥 < 2)
8 𝑦
Find the marginal p.d.f’s of X and Y. Are X and Y independent?
3 2 2 2
1 𝑥 𝑥 Solution:
∴ ℎ(𝑦) = [[ ] − 𝑦 [ ] ]
8 3 −𝑦 2 −𝑦
Marginal densities of X and Y:
1 8 𝑦3 𝑦3 ∞
= [ + − 2𝑦 + ] 𝑔(𝑥) = ∫ 𝑓 (𝑥, 𝑦)𝑑𝑦
8 3 3 2 −∞

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 13
 ∞ −(𝑥2 +𝑦2 ) 𝜕2
= ∫ 𝑥𝑦𝑒 2 𝑑𝑦 ∴ 𝑓 (𝑥, 𝑦) = [1 − 𝑒 −𝑥 − 𝑒 −𝑦 + 𝑒 −(𝑥+𝑦) ]
0 𝜕𝑥𝜕𝑦
−𝑥2 ∞ −𝑦2 𝜕 −𝑦
= 𝑥𝑒 2 ∫ 𝑦𝑒 2 𝑑𝑦 = [𝑒 + 𝑒 −(𝑥+𝑦) ]
0 𝜕𝑥
𝑦2 = 𝑒 −(𝑥+𝑦)
Put 𝑡 = 2
. 𝑇ℎ𝑒𝑛 𝑑𝑡 = 𝑦𝑑𝑦 𝑎𝑛𝑑 0 < 𝑡 < ∞.
𝑒 −(𝑥+𝑦) , 𝑥 ≥ 0, 𝑦 ≥ 0
−𝑥2 ∞ The joint p.d.f 𝑓 (𝑥, 𝑦) = {
∴ 𝑔(𝑥) = 𝑥𝑒 2 ∫ 𝑒 −𝑡 𝑑𝑡 0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
0
Marginal densities of X and Y:
−𝑥2 −𝑥2
∞ ∞
= 𝑥𝑒 2 [−𝑒 −𝑡 ]∞
0 = 𝑥𝑒 2 [0 + 1]
𝑔(𝑥) = ∫ 𝑓(𝑥, 𝑦)𝑑𝑦 = ∫ 𝑒 −(𝑥+𝑦) 𝑑𝑦
−𝑥2 0 0
⇒ 𝑔(𝑥) = 𝑥𝑒 2 ;𝑥 ≥ 0
= 𝑒 −𝑥 [−𝑒 −𝑦 ]∞ −𝑥
0 = 𝑒 ,𝑥 > 0
Similarly the marginal density of Y is ∞ ∞
∞ ℎ(𝑦) = ∫ 𝑓 (𝑥, 𝑦)𝑑𝑥 = 𝑒 −𝑦 ∫ 𝑒 −𝑥 𝑑𝑥
ℎ(𝑦) = ∫ 𝑓 (𝑥, 𝑦)𝑑𝑥 −∞ 0
0
= 𝑒 −𝑦 [−𝑒 −𝑥 ]∞
0 = 𝑒 −𝑦 , 𝑦 > 0
−𝑦2 ∞ −𝑥2
⇒ ℎ (𝑦) = 𝑦𝑒 2 ∫ 𝑥𝑒 2 𝑑𝑥 Conditional density functions:
0
𝑓 (𝑥, 𝑦)
−𝑦2
𝑥2 𝑓 (𝑥/𝑦) = = 𝑒 −𝑥 , 𝑥 > 0
= 𝑦𝑒 2 ,𝑦 ≥ 0 𝑢𝑠𝑖𝑛𝑔 𝑡 = ℎ (𝑦)
2
−(𝑥2 +𝑦2 ) 𝑓 (𝑥, 𝑦)
𝑔(𝑥)ℎ(𝑦) = 𝑥𝑦𝑒 2 𝑓 (𝑦/𝑥) = = 𝑒 −𝑦 , 𝑦 > 0
𝑔(𝑥)
= 𝑓 (𝑥, 𝑦) ⇒ 𝑋𝑎𝑛𝑑 𝑌 𝑎𝑟𝑒 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡. Since 𝑔(𝑥)ℎ (𝑦) = 𝑒 −(𝑥+𝑦) = 𝑓 (𝑥, 𝑦), X and Y are independent.
Problem 13: 1 1
𝑃[𝑋 ≤ 1; 𝑌 ≤ 1] = ∫ ∫ 𝑒 −(𝑥+𝑦) 𝑑𝑦𝑑𝑥
The joint distribution function of two random variables X and Y is given by 0 0
−𝑥 −𝑦 −(𝑥+𝑦) 1 1
𝐹 (𝑥, 𝑦) = 1 − 𝑒 −𝑒 +𝑒 , 𝑥 > 0, 𝑦 > 0 𝑎𝑛𝑑 = 0, otherwise. Find
= (∫ 𝑒 −𝑥 𝑑𝑥) (∫ 𝑒 −𝑦 𝑑𝑦)
the marginal and conditional densities of X and Y. Are X and Y independent? 0 0

Also compute 𝑃[𝑋 ≤ 1, 𝑦 ≤ 1]. 1 2

= (1 − 𝑒 −1 )(1 − 𝑒 −1 ) = (1 − )
Solution: 𝑒

𝜕2 𝐹(𝑥,𝑦)
= 0.3996
The joint p.d.f 𝑓 (𝑥, 𝑦) = 𝜕𝑥𝜕𝑦

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 14
Problem 14: 𝑓 (𝑥, 𝑦) 6(1 − 𝑥 − 𝑦)
𝑓(𝑥/𝑦) = =
Two continuous random variables X and Y have the joint p.d.f 𝑓 (𝑥, 𝑦) = 𝑓𝑌 (𝑦) 3(1 − 𝑦)2

6(1 − 𝑥 − 𝑦), 𝑥 > 0, 𝑦 > 0, 0 < 𝑥 + 𝑦 < 1 2(1 − 𝑥 − 𝑦)

{ = , 0 < 𝑥, 𝑦 < 1 𝑎𝑛𝑑 0 < 𝑥 + 𝑦 < 1
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 (1 − 𝑦)2
Find the marginal distribution and conditional distributions of X and Y. 𝑓 (𝑥, 𝑦) 6(1 − 𝑥 − 𝑦)
𝑓(𝑦/𝑥) = =
Hence examine if X and Y are independent. 𝑓𝑋 (𝑥) 3(1 − 𝑥)2
2(1 − 𝑥 − 𝑦)
Solution: = , 0 < 𝑥, 𝑦 < 1 𝑎𝑛𝑑 0 < 𝑥 + 𝑦 < 1
(1 − 𝑥)2
Marginal densities of X and Y:
∞ X+y= 𝑁𝑜𝑤 𝑓𝑋 (𝑥)𝑓𝑌 (𝑦) = 9(1 − 𝑥)2 (1 − 𝑦)2
𝑓𝑋 (𝑥) = ∫ 𝑓 (𝑥, 𝑦)𝑑𝑦 1 ≠ 6(1 − 𝑥 − 𝑦)
−∞ 1-x
1−𝑥 𝑓𝑋 (𝑥)𝑓𝑌 (𝑦) ≠ 𝑓 (𝑥, 𝑦) ⇒ 𝑋 𝑎𝑛𝑑 𝑌 𝑎𝑟𝑒 𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡.
=∫ 6(1 − 𝑥 − 𝑦)𝑑𝑦 1-y
0
1−𝑥 Problem 15:The joint probability density function of a R.V X and Y is given by
= 6∫ [(1 − 𝑥) − 𝑦]𝑑𝑦
𝑥 3𝑦3
0 0 0 , 0 < 𝑥 < 2 ,0 < 𝑦 < 2
𝑓 (𝑥, 𝑦) = { 16 . Find marginal density function of X
1−𝑥
𝑦2 0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
= 6 [(1 − 𝑥)𝑦 − ]
2 0
and Y.
2
(1 − 𝑥)2
)
= 6 [(1 − 𝑥 − ] Solution:
2
The m.d.f of ‘𝑋’ is
∴ 𝑓𝑋 (𝑥) = 3(1 − 𝑥)2 , 0 < 𝑥 < 1 ∞
1−𝑦
𝑓𝑋 (𝑥) = 𝑓(𝑥) = ∫ 𝑓 (𝑥, 𝑦)𝑑𝑦
𝑓𝑌 (𝑦) = ∫ 6(1 − 𝑥 − 𝑦)𝑑𝑦
0 −∞
1−𝑥 2 𝑥 3𝑦3 𝑥 3𝑦3
[(1 − 𝑦) − 𝑥]𝑑𝑥 = ∫0 𝑑𝑦 ∵ 𝑓(𝑥, 𝑦) = &0 < 𝑦 < 2
= 6∫ 16 16
0
1 3 2 3 𝑑𝑦 𝑥3
(1 − 𝑦)2 = 𝑥 ∫ 𝑦 = [𝑦 4 ]20
)2
= 6 [(1 − 𝑦 − ] 16 0 16 × 4
2
𝑥3 𝑥3
∴ 𝑓𝑌 (𝑦) = 3(1 − 𝑦)2 , 0 < 𝑦 < 1 = × 16 = ,0 < 𝑥 < 2
16 × 4 4
Conditional densities 𝑓(𝑥/𝑦) 𝑎𝑛𝑑 𝑓(𝑦/𝑥): The m.d.f of ‘Y’ is

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 15
 ∞ ∞
(ii) m.d.f of ‘𝑋’ is 𝑓(𝑥) = ∫−∞ 𝑓(𝑥, 𝑦)𝑑𝑦
𝑓𝑌 (𝑦) = 𝑓(𝑦) = ∫ 𝑓(𝑥, 𝑦) 𝑑𝑥
28
−∞ = ∫𝑥 9 𝑥𝑦𝑑𝑦 ∵ 𝑥 < 𝑦 < 2
2 𝑥 3𝑦3 𝑦3 2 𝑦3 𝑦3 8 2 8
= ∫0 𝑑𝑥 = ∫ 𝑥 3 𝑑𝑥 = [𝑥 4 ]20 = × 16 = 𝑥 ∫𝑥 𝑦𝑑𝑦 = 𝑥[𝑦 2 ]2𝑥
16 16 0 16×4 16×4 9 9×2
𝑦3 4
= ,0 < 𝑦 < 2 = 𝑥 (4 − 𝑥 2 ), 1 < 𝑥 < 2
4 9
∞ 𝑦8
Problem 16: Let X and Y be random variable with the joint pdff(x) (iii) m.d.f of‘𝑌’ is 𝑓(𝑦) = ∫−∞ 𝑓(𝑥, 𝑦)𝑑𝑥 = ∫1 9 𝑥𝑦𝑑𝑥
8
𝑥𝑦, 1 < 𝑥 < 𝑦 < 2 8 4 4
={𝑘 (i)Find K, (ii)𝑓𝑥 (𝑥), (iii)𝑓𝑦 (𝑌) and (iv) 𝑓(𝑋/𝑌). = 𝑦[𝑥 2 ]1𝑦 = 𝑦(𝑦 2 − 1) = y(y 2 − 1), 1 < 𝑦 < 2
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 9×2 9 9
Solution: (iv)Conditional density function of X given Y = y is
∞ ∞ 8
(i) Given 𝑓(𝑥, 𝑦) is joint pdf i.e.∫−∞ ∫−∞ 𝑓 (𝑥, 𝑦) 𝑑𝑦 𝑑𝑥 = 1 𝑓(𝑥, 𝑦) 𝑥𝑦 2𝑥
9
2 𝑥 (v)𝑓(𝑥/𝑦) = = 4𝑦 = ,1 < 𝑥 < 𝑦 < 2
8 𝑓(𝑦) (𝑦 2 − 1) 𝑦2 −1
∫ ∫ 𝑥𝑦 𝑑𝑦 𝑑𝑥 = 1 9
𝑘
1 1 Problem 17: The joint p.d.f of a bivariate R.V(x,y) is given by
𝑥
8 2 𝑥
⟹ 𝑘 ∫1 𝑥[∫1 𝑦𝑑𝑦] 𝑑𝑥 = 1
8 2
⟹ 𝑘 ∫1 𝑥 [2 (𝑦 2 )]
1
𝑑𝑥 = 1 𝑘𝑥𝑦, 0 < 𝑥 < 1,0 < 𝑦 < 1
𝑓(𝑥, 𝑦) = {
1 0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
2 2
8 𝑥 4 Find k, (ii) Find 𝑝(𝑥 + 𝑦 < 1), (iii) Are x and y independent R.V’s
⟹ ∫ (𝑥 2 − 1)𝑑𝑥 = 1 ⟹ ∫(𝑥 3 − 𝑥) 𝑑𝑥 = 1
𝑘 2 𝑘 Solution:
1 1
2 (i) Since 𝑓(𝑥, 𝑦) is a p.d.f., we have
4 𝑥4 𝑥2
⟹ { − } = 1 ∞ ∞
𝑘 4 2 1
4 ∫ ∫ 𝑓(𝑥, 𝑦) 𝑑𝑥 𝑑𝑦 = 1
⟹ { (4 − 2) – (1/(4 ) − 1/2 )} = 1 −∞ −∞
𝐾
4 1 1 1
⟹ 𝐾 [2+4] = 1 i.e.,∫0 ∫0 𝑘𝑥𝑦 𝑑𝑥 𝑑𝑦 = 1 𝑓 (𝑥, 𝑦) = 𝑘𝑥𝑦, [0 < 𝑥 < 1,0 < 𝑦 < 1]
1 1
4 9
⟹ ( )= 1∴𝐾=9 ⟹ 𝑘 ∫ 𝑦[∫ 𝑥 𝑑𝑥] 𝑑𝑦 = 1
𝐾 4
8 0 0
𝑥𝑦, 1 < 𝑥 < 𝑦 < 2
∴P.d.f of(𝑋, 𝑌) is 𝑓(𝑥, 𝑦) = {9 1 1
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 𝑘 𝑘
⟹ ∫ 𝑦 [𝑥 2 ]10 𝑑𝑦 = 1 ⟹ ∫ 𝑦 𝑑𝑦 = 1
2 2
0 0

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 16
 𝑘 21 1
⟹ [𝑦 ]0 = 1
4 = ∫ 4𝑥𝑦𝑑𝑦 , 0<𝑦<1
𝑘 0
⟹ =1 ⟹𝑘=4
4 𝑦²
= 4𝑥[ ]10 = 2𝑥, 0 < 𝑥 < 1 − − − − − − − (1)
4𝑥𝑦, 0 < 𝑥, 𝑦 < 1 2
Hence , the p.d.f of (𝑥, 𝑦) is 𝑓(𝑥, 𝑦) = {
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 The marginal density function of ‘𝑌’ is
(ii)To evaluate𝑝(𝑥 + 𝑦 < 1) ∞

1 1−𝑦 𝑓𝑌 (𝑦) = ∫ 𝑓 (𝑥, 𝑦)𝑑𝑥 (formula)

𝑃(𝑥 + 𝑦 < 1) = ∫ ∫ 𝑓(𝑥, 𝑦) 𝑑𝑥 𝑑𝑦 −∞

0 0 1
𝑥2
1 1−𝑦 1 1−𝑦 = ∫ 4𝑥𝑦𝑑𝑥 , 0 < 𝑥 < 1 = 4𝑦[ ]10
2
= 4 ∫ ∫ 𝑥𝑦 𝑑𝑥 𝑑𝑦 = 4 ∫ 𝑦[ ∫ 𝑥𝑑𝑥 ] 𝑑𝑦 0

0 0 0 0 = 2𝑦, 0 < 𝑦 < 1 − − − − − − − −(2)

1 1
4 Next, to check the two random variables X and Y are independent.
= ∫ 𝑦[𝑥 2 ]1−𝑦
0 𝑑𝑦 = 2 ∫ 𝑦(1 − 𝑦)²𝑑𝑦
2 Using formula, 𝑓𝑋 (𝑥) 𝑓𝑌 (𝑦) = 𝑓𝑋𝑌 (𝑥, 𝑦)
0 0
1 𝑓(𝑥) 𝑓(𝑦) = (2𝑥) (2𝑦) = 4𝑥𝑦 = 𝑓(𝑥, 𝑦)
= 2 ∫ 𝑦[1 − 2𝑦 + 𝑦 2 ]𝑑𝑦 Therefore, X and Y are independent.
0 Problem 18:If the joint P.d.f X and Y is given by 𝑓(𝑥, 𝑦) =
1
1
𝑦² 2𝑦³ 𝑦⁴ 1 (3𝑥 2 + 𝑥𝑦), 0 < 𝑥 < 1,0 < 𝑦 < 2
= 2 ∫[𝑦 − 2𝑦 2 + 𝑦 3 ]𝑑𝑦 = 2[ − + ]0 {3
2 3 4 0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
0
1 2 1 3 2 9−8 Find 𝑃(𝑋 + 𝑌 ≥ 1)
= 2[ − + ] = 2[ − ] = 2[ ] = 1/6.
2 3 4 4 3 12
Solution:
(ii) Next to check that X and Y are independent R.V’S
𝑃 (𝑋 + 𝑌 ≥ 1) = 1 − 𝑃 (𝑋 + 𝑌 < 1) − − − − − − − − − −(1)
First we find the marginal density functions X and Y.
1 1−𝑦
Now, 𝑃(𝑋 + 𝑌 < 1) = ∫0 ∫0 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦
The marginal density function of ‘X’ is
1 1−𝑦
∞ 1
=∫ ∫ (3𝑥 2 + 𝑥𝑦)𝑑𝑥𝑑𝑦
𝑓𝑋 (𝑥) = ∫ 𝑓 (𝑥, 𝑦)𝑑𝑦 (formula) 0 0 3
−∞ 1 1−𝑦
1
= ∫ (∫ (3𝑥 2 + 𝑥𝑦)𝑑𝑥 ) 𝑑𝑦
3 0 0

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 17
 1−𝑦
1 1 𝑥2 𝑦 1. The two dimensional random variable (X, Y) has the joint probability
= ∫ (𝑥 3 + ) 𝑑𝑦 𝑥+2𝑦
3 0 2 0 mass function 𝑓 (𝑥, 𝑦) = , 𝑥 = 0,1,2; 𝑦 = 0,1,2. Find the conditional
27
1−𝑦
1 1 (1 − 𝑦)2 𝑦 distribution of Y for X = x. Also find the conditional distribution of Y given
= ∫ ((1 − 𝑦)3 + ) 𝑑𝑦
3 0 2 0 X = 1.
2. For the bivariate probability distribution of (X, Y) given below, find
y X 1 2 3 4 5 6
1
(𝑖) 𝑃 (𝑋 ≤ 1)(𝑖𝑖)𝑃(𝑌 ≤ 3)(𝑖𝑖𝑖)𝑃(𝑋 ≤ 1, 𝑦 ≤ 3)(𝑖𝑣)𝑃 (𝑋 ≤ 𝑌 ≤ 3) (𝑣)𝑃 (𝑌 ≤
0 0 0 1⁄ 2⁄ 2⁄ 3⁄ 3 3
32 32 32 32 ≤ 1) (𝑣𝑖 )𝑃(𝑋 + 𝑌 ≤ 4)(𝑣)𝑃 (𝑌 ≤ ≤ 1) (𝑣𝑖)𝑃(𝑋 + 𝑌 ≤ 4)
𝑋 𝑋
1 1⁄ 1⁄ 1⁄ 1⁄ 1⁄ 1⁄ 7 23
16 16 8 8 8 8 ANS:(𝑖) 𝑃(𝑋 ≤ 1) = ; (𝑖𝑖 )𝑃(𝑌 ≤ 3) = ; (𝑖𝑖𝑖)𝑃(𝑋 ≤ 1, 𝑦 ≤ 3) =
8 64
2 1⁄ 1⁄ 1⁄ 1⁄ 0 2⁄ 9 1 18 3 9
32 32 64 64 64 ; (𝑖𝑣)𝑃 (𝑋 ≤ 𝑌 ≤ 3) = 23 ; (𝑣)𝑃 (𝑌 ≤ 𝑋 ≤ 1) = 28 ; (𝑣𝑖 )𝑃(𝑋 + 𝑌 ≤ 4) =
32
1 1 (1 − 𝑦)2 𝑦 13
= ∫ ((1 − 𝑦)3 + ) 𝑑𝑦 32
;
3 0 2
COVARIANCE:
1 1 1
= ∫ [1 − 𝑦 3 − 3𝑦 + 3𝑦 2 + (1 + 𝑦 2 − 2𝑦)𝑦] 𝑑𝑦 If X and Y are random variables, then covariance between X and Y is defined
3 0 2
1 1 𝑦 𝑦3 as
3 2
= ∫ [1 − 𝑦 − 3𝑦 + 3𝑦 + + − 2𝑦 2 ] 𝑑𝑦
3 0 2 2 𝐶𝑜𝑣 (𝑋, 𝑌) = 𝐸(𝑋𝑌) − 𝐸(𝑋). 𝐸(𝑌)
1 Formula
1 𝑦4 𝑦2 𝑦3 𝑦2 𝑦4 𝑦3
= [𝑦 − −3 +3 + + −2 ]
3 4 2 3 4 8 3 0 If X and Y are independent, then𝐸 (𝑋𝑌) = 𝐸(𝑋). 𝐸(𝑌)and𝐶𝑜𝑣(𝑋, 𝑌) = 0
1 1 3 1 1 2 1 1 3 2 (i) 𝐶𝑜𝑣(𝑎𝑋, 𝑏𝑌) = 𝑎𝑏 𝑐𝑜𝑣(𝑋, 𝑌)
= [1 − − + 1 + + − ] = [2 + − − ]
3 4 2 4 8 3 3 8 2 3 (ii) 𝐶𝑜𝑣(𝑎𝑋 + 𝑏, 𝑐𝑌 + 𝑏) = 𝑎𝑐 𝐶𝑜𝑣(𝑋, 𝑌)
1 17 13 1 102 − 104
= [ − ] = [ ] (iii) If X &Y are independent, then 𝑉𝑎𝑟(𝑋 ± 𝑌) = 𝑉𝑎𝑟 (𝑋) + 𝑉𝑎𝑟 (𝑌)
3 8 6 3 48
(iv) 𝐶𝑜𝑣(𝑋, 𝑋) = 𝑉𝑎𝑟(𝑋) = 𝜎𝑥2
1 −2 −1
= [ ]= −−−−−−−② (v) 𝐶𝑜𝑣(𝑋 + 𝑌, 𝑋 − 𝑌) = 𝜎𝑥2 − 𝜎𝑦2 .
3 48 72
7 73
𝑆𝑢𝑏② 𝑖𝑛 ①, 𝑤𝑒𝑔𝑒𝑡 𝑃[𝑋 + 𝑌 ≥ 1] = 1 + =
72 72
Problem for Practice:

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 18
CORRELATION: Here 𝑈 = 𝑥 − 𝑎, 𝑉 = 𝑦 − 𝑏 , n – number of items, Where a,b are
A distribution involving two variables is known as a bivariate distribution. If arbitrary values.
these two variables vary such that change in one variable affects the change Note:
in the other variables, the variable are said to be correlation. Correlation coefficient is independent of change of origin and scale.
TYPE OF CORRELATION 𝑋−𝑎 𝑌−𝑏
i.e.,𝑟(𝑋, 𝑌) = 𝑟(𝑈, 𝑉), 𝑤ℎ𝑒𝑟𝑒 𝑈 = ℎ
;𝑉 = 𝑘
(i) Positive Correlation (ii) Negative Correlation
Where a and b are some arbitrary constants usually the mid-values of the
POSITIVE CORRELATION
given data X and Y respectively.
Problems Based on Correlation Coefficient
X 65 67 66 71 67 70 68 69
Problem: 1 Compute the coefficients of correlation between X and Y using
Y 67 68 68 70 64 67 72 70
the following data.
If the two variables deviate in the same direction [i.e if the increase (or
Solution: Let us take U = X – 68 &V = Y − 68
decrease) in one results in a corresponding increase (or decrease) in the
U = X-
other, then the correlation is said to be positive. X Y V=Y-68 UV U2 V2
68
EXAMPLE: The correlation between (i) The height and weight of a group of
65 67 -3 -1 3 9 1
persons (ii) Income and expenditure
67 68 -1 0 0 1 0
NEGATIVE CORRELATION
If the two variables constantly deviate in opposite direction (i.e if the 66 68 -2 0 0 4 0

increase (or decrease) in one results in corresponding decrease (or 71 70 3 2 6 9 4

increase) in the other, then the correlation is said to be Negative. 67 64 -1 -4 4 1 16
KARL PEARSON’S COEFFICIENT OF CORRELATION 70 67 2 -1 -2 4 1
Formula: 68 72 0 4 0 0 16
If X and Y take considerably large values, computation of rxy will become 69 70 1 2 2 1 4
difficult. In such problems, we may introduce change of origin and scale and Total ΣU=-1 ΣV=2 ΣUV=13 ΣU2=29 ΣV2=42
compute rxy by using following property.
Cov(U,V) 𝑛𝛴𝑈𝑉 – 𝛴𝑈.𝛴𝑉
𝑟 (𝑥, 𝑦) = 𝑟𝑢𝑣 = =
σu σv √𝑛𝛴𝑈 2 – (𝛴𝑈)2√ 𝑛𝛴𝑉 2 – (𝛴𝑉)2

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 19
Using Formula X Y 𝐔 = X − 15 V = Y − 23 UV U2 V2

𝑛𝛴𝑈𝑉 − 𝛴𝑈. 𝛴𝑉 5 16 -10 -7 14 4 49

𝑟𝑥𝑦 = 𝑟𝑢𝑣 = , ℎ𝑒𝑟𝑒 𝑛 = 8
√(𝑛𝛴𝑈 2 − (𝛴𝑈 2 )). (𝑛𝛴𝑉 2 − (𝛴𝑉)2 ) 10 19 -5 -4 4 1 16
15 23 0 0 0 0 0
U = X- V=Y- 20 26 5 3 3 1 9
X Y UV U2 V2
69 69 25 30 10 7 14 4 49
65 67 -4 -2 8 16 4 Total ΣU=0 ΣV= -1 ΣUV=35 ΣU2=10 ΣV2=123
66 68 -3 -1 3 9 1 Using Formula,
67 65 -2 -4 8 4 16 𝑛𝛴𝑈𝑉 − 𝛴𝑈. 𝛴𝑉
𝑟𝑥𝑦 = 𝑟𝑢𝑣 ,𝑛 = 5
67 68 -2 -1 2 4 1 √(𝑛𝛴𝑈 2 − (𝛴𝑈)2 ). (𝑛𝛴𝑉 2 − (𝛴𝑉)2 )
68 72 -1 3 -3 1 9 5 × 35 − (0) × (−1) 175
= = = 0.9988
69 72 0 3 0 0 9 √(5 × 10 − 0). (5 × 123 − (−1)2 ) √50 × 614 ;

70 69 1 0 0 1 0
72 71 3 2 6 9 4 Problem: 3 calculate the correlation coefficient for the following heights (in

Total ΣU=-8 ΣV=0 ΣUV=24 ΣU2=44 ΣV2=44 inches) of fathers X and Y and their sons Y.

8 × 13 − (−1) × (2) 106

= = = 0.3828
√(8 × 29 − (−1)2 ). (8 × 42 − (2)2 √231 × 332 X 65 66 66 67 68 69 70 72
Y 67 68 65 68 72 72 69 71

Problem: 2 Find the coefficient of correlation between X and Y using the

following data Solution:

X 5 10 15 20 25 Let us take U = X – 69 : V = Y – 69

Y 16 19 23 26 30 Using Formula
𝑛𝛴𝑈𝑉 − 𝛴𝑈. 𝛴𝑉
𝑟𝑥𝑦 = 𝑟𝑢𝑣 = ℎ𝑒𝑟𝑒 𝑛 = 8
√(𝑛𝛴𝑈 2 − (𝛴𝑈)2 ). (𝑛𝛴𝑉 2 − (𝛴𝑉)2 )
Solution:
Let us take U = X − 15 ; V = Y – 23
Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 20
 8 × 24 − (−8) × (0) 7 × 48 − (−3) × (0) 336
= = = = 0.898
√(8 × 44 − (−8)2 ). (8 × 44 − (0)2 ) √(7 × 37 − (−3)2 ). (7 × 80 − (0)2 √250x560
192
= = 0.6030
√288√352
Problem: 4 Find the coefficient of correlation between industrial production
and export using the following data
Production X 55 56 58 59 60 60 62
Export Y 35 38 37 39 44 43 44

Solution:
To find 𝒓𝒙𝒚 = 𝒓𝒖𝒗
Let us take U = X – 59 ; V = Y – 40
x y U = X-59 V=Y-40 UV U2 V2
55 35 -4 -5 20 16 25
56 38 -3 -2 6 9 4
58 37 -1 -3 3 1 9
59 39 0 -1 0 0 1
60 44 1 4 4 1 16
60 43 1 3 3 1 9
62 44 3 4 12 9 16
Total ΣU=-3 ΣV=0 ΣUV=48 ΣU2=37 ΣV2=80

Using Formula,
𝑛𝛴𝑈𝑉 − 𝛴𝑈. 𝛴𝑉
𝑟𝑥𝑦 = 𝑟𝑢𝑣
√(𝑛𝛴𝑈 2 − 𝛴𝑈 2 ). (𝑛𝛴𝑉 2 − 𝛴𝑉)2

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 21
Two dimensional Random variable : Moments - Mean ,Variance ,variance
,Standard variance PROBLEMS USING DISCRETE RANDOM VARIABLES:
Discrete Continuous Problem: 1 If X and Y are discrete R.V’s with p.d.f
𝑀𝑒𝑎𝑛 𝑜𝑓 𝑋 = 𝐸 (𝑋) 𝑀𝑒𝑎𝑛 𝑜𝑓 𝑋 𝐸(𝑋) 𝑥+𝑦
𝑓 (𝑥, 𝑦) = , 𝑥 = 1,2,3: 𝑦 = 1,2.
∞ 21
3
= ∫ 𝑥𝑓𝑋 (𝑥) 𝑑𝑥 Find (𝑖 )𝑀𝑒𝑎𝑛 𝑎𝑛𝑑 𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 𝑜𝑓 𝑥 𝑎𝑛𝑑 𝑦, (𝑖𝑖 )𝐶𝑜𝑣(𝑥, 𝑦), (𝑖𝑖𝑖 )𝑟(𝑥, 𝑦).
= ∑ 𝑥. 𝑓 (𝑥 )
−∞ 𝑥+𝑦
𝑥=1 Solution: Giventhat 𝑓(𝑥, 𝑦) = 21
, 𝑥 = 1, 2,3; 𝑦 = 1,2
3
𝐸(𝑋 2 ) (i) To find Mean, 𝑉𝑎𝑟(𝑋) and 𝑉𝑎𝑟(𝑌),
𝐸(𝑋 2 ) = ∑ 𝑥 2 . 𝑓(𝑥 ) ∞
we need m.d.f. of X, m.d.f. of 𝑌, 𝐸(𝑋), 𝐸(𝑋 2 ), 𝐸(𝑌) 𝑎𝑛𝑑 𝐸(𝑌 2 )
𝑥=1 = ∫ 𝑥 2 𝑓𝑋 (𝑥) 𝑑𝑥
−∞ The Marginal density function of X and Y is given by
2 2 2 2
𝑉𝑎𝑟(𝑋) = 𝐸[𝑋 ] − [𝐸(𝑋)] 𝑉𝑎𝑟(𝑋) = 𝐸[𝑋 ] − [𝐸(𝑋)]
1 2 3
Y X P(Y = y) or f(y)
𝑆 .𝐷 𝜎𝑥 = √(𝑉𝑎𝑟(𝑋) 𝑆 .𝐷 𝜎𝑥 = √(𝑉𝑎𝑟(𝑋)
1 2/21 3/21 4/21 9/21
3 2 𝐸(𝑋𝑌)
𝐸 (𝑋𝑌) = ∑ ∑(𝑥𝑦) 𝑓 (𝑥, 𝑦) 2 3/21 4/21 5/21 12/21
∞ ∞
𝑥=1 𝑦=1 = ∫ ∫ 𝑥𝑦𝑓 (𝑥, 𝑦)𝑑𝑥𝑑𝑦 P(X = x) or f(x) 5/21 7/21 9/21 1
−∞ −∞

𝐶𝑜𝑣(𝑋, 𝑌) = 𝐸(𝑋𝑌) 𝐶𝑜𝑣 (𝑋, 𝑌)

Mean of 𝑋 = 𝐸(𝑋) = ∑3𝑥=1 𝑥. 𝑓(𝑥) = (1). 𝑓 (1) + 2. 𝑓 (2) + 3. 𝑓(3)
− 𝐸(𝑋) . 𝐸(𝑌) = 𝐸(𝑋𝑌)
5 7 9 5 + 14 + 27 46
= + 2( )+ 3( ) = =
− 𝐸(𝑋) 𝐸(𝑌) 21 21 21 21 21
3
𝐶𝑜𝑣(𝑋, 𝑌) 𝐶𝑜𝑣(𝑋, 𝑌)
𝑟(𝑋, 𝑌) = 𝑟(𝑋, 𝑌) = 𝐸(𝑋 ) = ∑ 𝑥 2 . 𝑓 (𝑥) = (1)2 . 𝑓(1) + 22 𝑓(2) + 32 𝑓(3)
2
𝜎𝑥 𝜎𝑦 𝜎𝑥 𝜎𝑦
𝑥=1

5 7 9 5 + 28 + 81 114
= 1( ) + 4 ( ) + 9 ( ) = =
21 21 21 21 21
Mean of 𝑌 = 𝐸(𝑌) = ∑2𝑦=1 𝑦. 𝑓(𝑦) = (1). 𝑓(1) + 2. 𝑓(2). 𝑓(2)
9 12 9 + 24 33
= + 2( ) = =
21 21 21 21

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 22
 2

𝐸(𝑌 ) = ∑ 𝑦 2 . 𝑓 (𝑦) = (1)2 . 𝑓 (1) + 22 . 𝑓(2)

2
𝐶𝑜𝑣(𝑋, 𝑌) −6/441
𝑦=1 (iii) 𝑟(𝑋, 𝑌) = = = −0.0346
𝜎𝑥 𝜎𝑦 278 108
9 12 9 + 48 57 √ √
441 441
= 1( ) + 4 ( ) = =
21 21 21 21
𝑉𝑎𝑟(𝑋) = 𝐸[𝑋 2 ] − [𝐸(𝑋)]2
Problem: 2 The following table gives the joint probability distribution of two
114 46 2 114 2116 2394 − 2116 278
= −( ) = − = = ; random variable X and Y. Find 𝐸(𝑋), 𝐸(𝑌) 𝑎𝑛𝑑 𝐸(𝑋𝑌). Verify whether X and
21 21 21 441 441 441
Y are correlated.
√278
𝜎𝑥 = √(𝑉𝑎𝑟(𝑋) = √278/441 = Y/X 0 1 2 3
21
2 2 1 1 1 1
𝑉𝑎𝑟(𝑌) = 𝐸[𝑌 ] − [𝐸(𝑌)]
2
57 33 57 1089 1197 − 1089 108 8 8 8 8
= − ( )2 = − = = ; 1 1 1
21 21 21 441 441 441 3 0
16 8 16
108 √108 1 1 1
𝜎𝑦 = √(𝑉𝑎𝑟(𝑌) = √ = 4 0
441 21 16 8 16
(ii ) To find 𝐶𝑜𝑣(𝑋, 𝑌) Solution
We know that 𝐶𝑜𝑣(𝑋, 𝑌) = 𝐸(𝑋𝑌) − 𝐸(𝑋) . 𝐸(𝑌) −----------------------- (1) The m.d.f. of X and Y is given by
Now, 𝐸(𝑋𝑌) = ∑3𝑥=1 ∑2𝑦=1(𝑥𝑦) 𝑓(𝑥, 𝑦) Y/X 0 1 2 3 P(Y = y)
= (1)(1)𝑓 (1,1) + (1)(2)𝑓 (1,2) + (2)(1)𝑓 (2,1) + (2)(2)𝑓(2,2) 1 1 1 1 1
2
8 8 8 8 2
+ (3)(1)𝑓(3,1) + (3)(2)𝑓(3,2)
1 1 1 1
2 3 3 4 4 5 3 0
= + 2 ( ) + 2 ( ) + 4 ( ) + 3 ( ) + 6( ); 16 8 16 4
21 21 21 21 21 21
1 1 1 1
2 + 6 + 6 + 16 + 12 + 30 4 0
= = 72/21 16 8 16 4
21
1 1 1 1
(1) Becomes, P(X = x) 𝟏
4 4 4 4
72 46 33 1512 − 1518 −6
𝐶𝑜𝑣(𝑋, 𝑌) = − ( )( ) = = ;
21 21 21 441 441
To find 𝐸(𝑋), 𝐸(𝑌) 𝑎𝑛𝑑𝐸(𝑋𝑌).

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 23
𝐸(𝑋) = ∑ 𝑥𝑝(𝑥), 𝐸(𝑦) = ∑ 𝑦𝑝(𝑦)𝑎𝑛𝑑𝐸(𝑋𝑌) = ∑ ∑(𝑥𝑦)𝑝(𝑥, 𝑦) Problem: 3 If the independent random variables X and Y have the variances

3
36 and 16 respectively. Find the correlation coefficient between X + Y and

𝑖. 𝑒. , 𝐸(𝑋) = ∑ 𝑥𝑝(𝑥) = 0. 𝑝(0) + 1. 𝑝(1) + 2. 𝑝(2) + 3. 𝑝(3) X - Y.

𝑥=0 Solution:Let 𝑈 = 𝑋 + 𝑌; 𝑉=𝑋−𝑌
1 1 1 6
=0+ +2( ) + 3( ) = We know that
4 4 4 4
4
𝐶𝑜𝑣(𝑈, 𝑉) 𝐸 (𝑈𝑉) − 𝐸(𝑈). 𝐸(𝑉)
𝑟𝑥𝑦 = 𝑟𝑢𝑣 = = − − − (1)
𝑖. 𝑒, . 𝐸(𝑌) = ∑ 𝑦𝑝(𝑦) = 2. 𝑝(2) + 3. 𝑝(3) + 4. 𝑝(4) √𝜎𝑢2 . 𝜎𝑣2 √[𝐸 (𝑈 2 ) − 𝐸(𝑈)2 ]. [𝐸 (𝑉 2 ) − 𝐸(𝑉)2 ]]
𝑦=2 Now, 𝐸(𝑈) = 𝐸(𝑋) + 𝐸(𝑌); 𝐸(𝑉) = 𝐸(𝑋) − 𝐸(𝑌)
1 1 1 11
= 2( ) + 3( )+ 4( ) = 𝐸(𝑈𝑉) = 𝐸(𝑋 2 − 𝑌 2 ) = 𝐸(𝑋 2 ) − 𝐸(𝑌 2 )
2 4 4 4
4 3 𝐸(𝑈 2 ) = 𝐸[(𝑋 + 𝑌)2 ] = 𝐸(𝑋 2 + 𝑌 2 + 2𝑋𝑌) = 𝐸(𝑋 2 ) + 𝐸(𝑌 2 ) + 2𝐸(𝑋𝑌)
𝐸(𝑋𝑌) = ∑ ∑(𝑥𝑦) 𝑝(𝑥, 𝑦) 𝐸(𝑉 2 ) = 𝐸(𝑋 2 ) + 𝐸(𝑌 2 ) − 2𝐸(𝑋𝑌)
𝑗=2 𝑖=0
Now, 𝐶𝑜𝑣(𝑈, 𝑉) = 𝐸(𝑈𝑉) − 𝐸(𝑈). 𝐸(𝑉)
= 0 × 2. p(0,2) + 0 × 3. p(0,3) + 0 × 4. p(0,4) + 1 × 2. p(1,2) + 1
= 𝐸(𝑋 2 ) − 𝐸(𝑌 2 ) − [(𝐸(𝑋) + 𝐸(𝑌)) . (𝐸(𝑋) − 𝐸(𝑌))]
× 3. p(1,3) × 4. p(1,4) + 2 × 2. p(2,2) + 2 × 3. p(2,3) + 2
= 𝐸(𝑋 2 ) − 𝐸(𝑌 2 ) − [𝐸 2 (𝑋 ) − 𝐸 2 (𝑌)]
× 4. p(2,4) + 3 × 2 . p(3,2) + 3 × 3. p(3,3) + 3 × 4 . p(3,4)
= [𝐸(𝑋 2 ) − 𝐸 2 (𝑋)] − [𝐸(𝑌 2 ) − 𝐸 2 (𝑌)]
1 1 1 1
= 0 + 0 + 0 + 2 ( ) + 3 ( ) + 4(0) + 4 ( ) + 6(0) + 8 ( ) = 𝜎𝑋2 − 𝜎𝑌2 = 36 − 16 = 20
8 8 8 8
+ 6(1/8) + 9(1/16) + 12(1/16) 𝜎𝑢2 = 𝐸(𝑈 2 ) − 𝐸 2 (𝑈)

2 3 4 8 6 9 12 4 + 6 + 8 + 16 + 12 + 9 + 12 = [𝐸(𝑋 2 ) + 𝐸 (𝑌 2 ) + 2𝐸 (𝑋𝑌)] − {[𝐸 2 (𝑋)] + [𝐸 2 (𝑌)] + 2𝐸 (𝑋). 𝐸(𝑌)}

= + + + + + + =
8 8 8 8 8 16 16 16 = [𝐸(𝑋 2 ) − 𝐸 2 (𝑋)] + [ 𝐸(𝑌 2 ) − [𝐸 2 (𝑌)] + 2[𝐸(𝑋𝑌) − 𝐸(𝑋). 𝐸(𝑌)]
67 = 36 + 16 + (2 × 0) = 52[ X and Y are independent R.V’s ]
𝐸(𝑋𝑌) =
16
𝜎𝑣2 = 𝐸(𝑉 2 ) − 𝐸 2 (𝑉)
67 3 11 67 33 1
𝐶𝑜𝑣(𝑋, 𝑌) = 𝐸 (𝑋𝑌 ) − 𝐸(𝑋). 𝐸(𝑌) = − . = − = = [𝐸(𝑋 2 ) + 𝐸(𝑌 2 ) − 2𝐸 (𝑋𝑌)] − {[𝐸 2 (𝑋)] + [𝐸 2 (𝑌)] − 2𝐸 (𝑋). 𝐸(𝑌)}
16 2 4 16 8 16
𝐶𝑜𝑣 (𝑋, 𝑌) ≠ 0 = [𝐸(𝑋 2 ) − 𝐸 2 (𝑋)] + [ 𝐸(𝑌 2 ) − [𝐸 2 (𝑌)] − 2[𝐸(𝑋𝑌) − 𝐸(𝑋). 𝐸(𝑌)]
Hence X and Y are correlated. = 36 + 16 − (2 × 0) = 52[ X and Y are independent R.V’s ]
20 20 5
(1) becomes, 𝑟𝑢𝑣 = √52.√52 = 52 = 13

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 24
 ∞
PROBLEMS USING CONTINOUS RANDOM VARIABLES: m.d.f. of Y is 𝑓𝑌 (𝑦) = ∫−∞ 𝑓(𝑥, 𝑦) 𝑑𝑥
Problem: 1 If the joint p.d.f of (𝑋, 𝑌) is given by 𝑓 (𝑥, 𝑦) = 2, 𝑖𝑛 1 1
𝑥2 1
0 ≤ 𝑥 < 𝑦 ≤ 1 . Find 𝐸(𝑋). = ∫ (2 − 𝑥 − 𝑦) 𝑑𝑥 = [(2 − 𝑦)𝑥 − ] = 2−𝑦−
0 2 0 2
Solution: 3
∞ 1 𝑓𝑌 (𝑦) = {2 − 𝑦 , 0≤𝑦≤1
𝑓 (𝑥) = ∫ 𝑓(𝑥 , 𝑦) = ∫ 2 𝑑𝑦 = (2𝑦)1𝑥 = 2(1 − 𝑥) 0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
−∞ 𝑥 To find 𝑣𝑎𝑟(𝑋 ) 𝑎𝑛𝑑 𝑣𝑎𝑟(𝑌), we need 𝐸(𝑋), 𝐸(𝑌), 𝐸(𝑋 2 ) 𝑎𝑛𝑑 𝐸(𝑌 2 )
∞
𝐸(𝑋) = ∫−∞ 𝑥 𝑓 (𝑥) 𝑑𝑥 ∞
Now, 𝐸(𝑋) = ∫−∞ 𝑥𝑓𝑋 (𝑥) 𝑑𝑥
1
Y=1
1 1
3 3
= ∫ 2 (𝑥 − 𝑥 2 ) 𝑑𝑥 = ∫ 𝑥( − 𝑥) 𝑑𝑥 = ∫ ( 𝑥 − 𝑥 2 ) 𝑑𝑥
0 2 0 2
0
X 1
1
= 3
𝑥2 𝑥3 1 X=Y = [ . 𝑥 2 − 𝑥 3 /3] = ¾ − 1/3 = (9 − 4)/12 = 5/12
= 2( − ) = 0 4 0
2 3 0 3
∞ 1 1
3 3 5
𝐸(𝑌) = ∫ 𝑦𝑓𝑌 (𝑦) 𝑑𝑦 = ∫ 𝑦( − 𝑦) 𝑑𝑦 = ∫ ( 𝑦 − 𝑦 2 ) 𝑑𝑦 =
−∞ 0 2 0 2 12
Problem: 2 Two random variables X and Y have the following joint ∞

probability density function 𝐸(𝑋 2 ) = ∫ 𝑥 2 𝑓𝑋 (𝑥) 𝑑𝑥

−∞
2 − 𝑥 − 𝑦, 0 ≤ 𝑥 ≤ 1, 0 ≤ 𝑦 ≤ 1 1 1
𝑓(𝑥, 𝑦) = { Find var(X), var(Y) & covariance 3 3
0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 = ∫ 𝑥 2 ( − 𝑥) 𝑑𝑥 = ∫ ( 𝑥 2 − 𝑥 3 ) 𝑑𝑥
0 2 0 2
between X & Y. Also find 𝑟𝑥𝑦
1
3 𝑥4 1 1 1
Solution: = [ × 𝑥3 − ] = − =
6 4 0 2 4 4
First we find the m.d.f of X and Y.
1
∞ Similarly, 𝐸 (𝑌 2 ) = 4
m.d.f of X is 𝑓𝑋 (𝑥) = ∫−∞ 𝑓(𝑥, 𝑦) 𝑑𝑦
1 i) 𝑉𝑎𝑟(𝑋 ) = 𝐸[𝑋 2 ] − [𝐸(𝑋)]2
𝑦2 1
= ∫ (2 − 𝑥 − 𝑦) 𝑑𝑦 = [(2 − 𝑥)𝑦 − ]10 = 2 − 𝑥 − 1 5 1 25 36 − 25 11
0 2 2 = − ( )2 = − = =
4 12 4 144 144 144
11 √11
3 𝜎𝑥2 = ⇒ 𝜎𝑥 =
144 12
𝑓𝑋 (𝑥) = {2 − 𝑥, 0 ≤ 𝑥 ≤ 1
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 25
 1 5 11 To find 𝑟𝑥𝑦 =
𝑐𝑜𝑣(𝑋,𝑌)
− − − − − − − −(1)
(𝑖𝑖) 𝑉𝑎𝑟(𝑌) = 𝐸[𝑌 2 ] − [𝐸(𝑌)]2 = − = 𝜎𝑥.𝜎𝑦
4 12 144
11 √11 First we find the m.d.f of X and Y and then find
𝜎𝑦2 = ⇒ 𝜎𝑦 =
144 12 𝐸(𝑋), 𝐸 (𝑌), 𝐸 (𝑋 2 ), 𝐸(𝑌 2 ), 𝜎𝑥 𝑎𝑛𝑑 𝜎𝑦
(iii) 𝑐𝑜𝑣 (𝑋, 𝑌) = 𝐸(𝑋𝑌) − 𝐸(𝑋) 𝐸(𝑌) m.d.f of X is𝑓𝑋 (𝑥)
∞
= ∫−∞ 𝑓(𝑥, 𝑦) 𝑑𝑦
∞ ∞
Now, 𝐸(𝑋𝑌) = ∫−∞ ∫−∞ 𝑥𝑦𝑓 (𝑥, 𝑦)𝑑𝑥𝑑𝑦 4 4
6−𝑥−𝑦 1 𝑦2
1 1 = ∫ 𝑑𝑦 = [(6 − 𝑥)𝑦 − ]
= ∫ ∫ 𝑥𝑦(2 − 𝑥 − 𝑦)𝑑𝑥𝑑𝑦 8 8 2 2
2
0 0
1 1
1 1 = 8 [(6 − 𝑥)4 − 8] – [8 [(6 − 𝑥)2 − 2]
= ∫ ∫ (2𝑥𝑦 − 𝑥 2 𝑦 − 𝑥𝑦 2 )𝑑𝑥𝑑𝑦
0 0 1 1
= [24 − 4𝑥 − 8 − 12 + 2𝑥 + 2] = [(6 − 2𝑥)] ,0 ≤ 𝑥 ≤ 2
1 1 1 8 8
𝑥2 𝑥3 𝑥2 𝑦 𝑦2
= ∫ [2 𝑦 − 𝑦 − 𝑦 2 ] 𝑑𝑦 = ∫ [𝑦 − − ] 𝑑𝑦 𝑚. 𝑑. 𝑓 𝑜𝑓 𝑌 𝑖𝑠 𝑓𝑌 (𝑦) = ∫−∞ 𝑓(𝑥, 𝑦) 𝑑𝑥 = ∫0
∞ 2 6−𝑥−𝑦
𝑑𝑥
0 2 3 2 0
3 2 8
0
2
1 1 𝑥2
𝑦2 𝑦2 𝑦3 1 1 1 1 2 1 = [(6 − 𝑦)𝑥 − ]
=[ − − ] = − − = − = 8 2 0
2 6 6 0 2 6 6 2 6 6
1 1
1 5 5 1 25 25 1 = [(6 − 𝑦)2 − 2] = [10 − 2𝑦] , 2 ≤ 𝑦 ≤ 4
𝑐𝑜𝑣 (𝑋, 𝑌) = −( ) = − = 24 – = − 8 8
6 12 12 6 144 144 144 ∞

𝐸(𝑋) = ∫ 𝑥𝑓(𝑥) 𝑑𝑥
1
𝑐𝑜𝑣(𝑋, 𝑌) 𝑐𝑜𝑣(𝑋, 𝑌) − 1 −∞
144
(𝑖𝑣)𝑟𝑥𝑦 = 𝑟(𝑋, 𝑌) = = = =− 2 2 2
𝜎𝑥 𝜎𝑦 √11
.
√11 11 1 1 1 𝑥2 2
√𝜎𝑥2 . √𝜎𝑦2 . 12 12 = ∫ 𝑥. [6 − 2𝑥] 𝑑𝑥 = ∫[6𝑥 − 2𝑥 2 ] 𝑑𝑥 = [6 − 𝑥 3 ]
8 8 8 2 3 0
0 0

1 1 36 − 16 20 5
Problem: 3 If 𝑓(𝑥, 𝑦) =
6−𝑥−𝑦
, 0 ≤ 𝑥 ≤ 2 , 2 ≤ 𝑦 ≤ 4 , find the correlation = [12 − 16/3] = [ ] = =
8 8 8 3 8×3 6
∞
coefficient between X & Y.
𝐸(𝑌) = ∫ 𝑦𝑓(𝑦)𝑑𝑦
Solution:
−∞
1
Given 𝑓(𝑥, 𝑦) = (6 − 𝑥 − 𝑦), 0 ≤ 𝑥 ≤ 2 , 2 ≤ 𝑦 ≤ 4
8

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 26
 4 4
1 1 1 𝑦2 2
4 1 4 2
= ∫ 𝑦 [10 − 2𝑦] 𝑑𝑦 = ∫[10𝑦 − 2𝑦 2 ] 𝑑𝑦 = [10 − 𝑦 3 ] = ∫ [∫ (6𝑥𝑦 − 𝑥 2 𝑦 − 𝑥𝑦 2 )𝑑𝑥] 𝑑𝑦
8 8 8 2 3 8 2 0
2 2 2
2
1 128 16 1 112 1 4 𝑥2 𝑦 𝑥3 𝑦 𝑥2 𝑦 2 1 4 8
= ∫ [6 − − ] 𝑑𝑦 = ∫ [12𝑦 − 𝑦 − 2𝑦 2 ] 𝑑𝑦
= [(20 − ) − (20 − )] = [60 − ] 8 2 2 3 2 0 8 2 3
8 3 3 8 3
1 180 − 112 1 17 1 12 2 8 2 2 3 4 1 64 128 16 16
= [ ] = × 68 = = [ 𝑦 − 𝑦 − 𝑦 ] = {(96 − − − 24 + + )}
8 3 24 6 8 2 6 3 2 8 3 3 3 3
∞
1 −64 − 128 + 16 + 16 1 −160
𝐸(𝑋 ) =2 ∫ 𝑥 2 𝑓(𝑥) 𝑑𝑥 = [72 + ( )] = [72 + ]
8 3 8 3
−∞
1 216 − 160 7
2 2 = [ ]=
1 1 8 3 3
= ∫ 𝑥 . [6 − 2𝑥] 𝑑𝑥 = ∫(6𝑥 2 − 2𝑥 3 ) 𝑑𝑥
2
8 8 Now, 𝑣𝑎𝑟(𝑋) = 𝐸(𝑋 2 ) − [𝐸(𝑋)]2
0 0
2 5 2 25 11
1 𝑥3 2 1 =1− ( ) = 1− =
= [6 − 𝑥 4 ] = [16 − 8] = 1 6 36 36
8 3 4 0
8
∞
√11 √11
𝜎𝑥 = √𝑣𝑎𝑟(𝑋) = =
36 6
𝐸(𝑌 2 ) = ∫ 𝑦 2 𝑓(𝑦) 𝑑𝑦
−∞
𝑣𝑎𝑟(𝑌) = 𝐸(𝑌 2 ) − [𝐸(𝑌)]2
4 4 25 17 2 25 289 300 − 289 11
1 1 = −( ) = − = =
= ∫ 𝑦 2 . [10 − 2𝑦] 𝑑𝑦 = ∫(10𝑦 2 − 2𝑦 3 ) 𝑑𝑦 3 6 3 36 36 36
8 8
2 2
√11 √11
3 4 𝜎𝑦 = √𝑣𝑎𝑟(𝑌) = =
1 𝑦 2 1 640 80 36 6
= [10 − 𝑦 4 ] = {( − 128) − ( − 8)}
8 3 4 2
8 3 3 Hence (1) becomes
1 560 1 560 − 360 200 25 𝐸 (𝑋𝑌) − 𝐸(𝑋). 𝐸(𝑌)
= { − 120} = [ ] = = 𝑟𝑥𝑦 =
8 3 8 3 24 3 𝜎𝑥 𝜎𝑦
7 5 17 7 85 84−85
3
− (6) ( 6 ) − 36 −1 36
∞ ∞ 3 36
𝐸(𝑋𝑌) = ∫−∞ ∫−∞ 𝑥𝑦. 𝑓 (𝑥, 𝑦)𝑑𝑥𝑑𝑦 = = 11 = 11 = ×
(
√11 √11
)( ) 36 11
6 6 36 36
4 2
1
= ∫ ∫ (𝑥𝑦). [6 − 𝑥 − 𝑦] 𝑑𝑥 𝑑𝑦 −1
2 0 8 𝑟𝑥𝑦 = .
11

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 27
 1
= 3 ∫ [(𝑥 + 𝑦) − (𝑥 2 + 𝑦 2 )] 𝑑𝑦
Problem: 4 The joint probability density function of two random variables X 0
1
𝑘[(𝑥 + 𝑦) − (𝑥 2 + 𝑦 2 )], 0 < (𝑥 , 𝑦) < 1} 𝑦2 𝑦3
and Y is given 𝑓(𝑥 , 𝑦) = { Show = 3 [𝑥𝑦 + − 𝑥2 𝑦 − ]
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 2 3 0
that X and Y are uncorrelated but not independent. 1 1
Solution: = 3 [𝑥 + − 𝑥2 − ]
2 3
(i) We know that the function f(x, y) should satisfy 1
𝑓 (𝑥) = 3 [𝑥 − 𝑥 2 + ] 0<𝑥<1
∞ ∞ 1 1 6
1
∫ ∫ 𝑓(𝑥, 𝑦)𝑑𝑦𝑑𝑥 = 1 ∫ ∫ 𝐾[(𝑥 + 𝑦) − (𝑥 2 + 𝑦 2 )] 𝑑𝑥 𝑑𝑦 = 1 Similarly𝑓(𝑦) = 3 [𝑦 − 𝑦 2 + 6] 0<𝑦<1
−∞ −∞ 0 0
∞
1
1 1 1
𝑥 2
𝑥 3 𝐸(𝑋) = ∫ 𝑥𝑓 (𝑥) 𝑑𝑥 = ∫ 𝑥 3 (𝑥 − 𝑥 2 + ) 𝑑𝑥
⟹ 𝐾 ∫ ( + 𝑥𝑦 − − 𝑥𝑦 2 ) 𝑑𝑦 = 1 0 6
2 3 0
−∞
0
1
𝑥
1
1
1 = ∫ 3 (𝑥 2 − 𝑥 3 + ) 𝑑𝑥
6
⟹ 𝐾 ∫ ( + 𝑦 − − 𝑦 2 ) 𝑑𝑦 = 1 0
2 3 1
0 𝑥3 𝑥4 𝑥2 1 1 1 4−3+1 1
1 =3 [ − + ] =3 [ − + ]=3 [ ]=
3 4 12 0 3 4 12 12 2
1
⟹ 𝐾 ∫ ( + 𝑦 − 𝑦 2 ) 𝑑𝑦 = 1 1 1
6 𝐸(𝑋) = Similarly 𝐸 (𝑌) =
0 2 2
3 1 ∞ ∞
𝑦 𝑦2 𝑦 𝐸(𝑋𝑌) = ∫−∞ ∫−∞ 𝑥𝑦. 𝑓 (𝑥, 𝑦)𝑑𝑥𝑑𝑦
⇒ 𝐾[ + − ] =1
6 2 3 0 1 1

1 1 1 1+3−2 𝐾 = 3 ∫ ∫ 𝑥𝑦 [(𝑥 + 𝑦) − (𝑥 2 + 𝑦 2 )] 𝑑𝑥 𝑑𝑦
⟹𝐾 [ + − ]=1⟹𝐾 [ ]= =1⟹𝐾=3
6 2 3 6 3 0 0
1 1

𝑐𝑜𝑣(𝑋,𝑌)
= 3 ∫ ∫[𝑥 2 𝑦 + 𝑥𝑦 2 − 𝑥 3 𝑦 − 𝑥𝑦 3 ] 𝑑𝑥 𝑑𝑦
If 𝑟𝑥𝑦 = = 0 then X and Y are uncorrelated. 0 0
𝜎𝑥.𝜎𝑦
1 1
First we find the m.d.f of X and Y and then find 𝐸(𝑋), 𝐸(𝑌), 𝐸(𝑋𝑌) 𝑥3𝑦 𝑥2 𝑦 2 𝑥4 𝑦 𝑥2 𝑦 3
= 3 ∫[ + − − ] 𝑑𝑦
∞
m.d.f of X is 𝑓𝑋 (𝑥) = ∫−∞ 𝑓(𝑥, 𝑦) 𝑑𝑦 3 2 4 2 0
0

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 28
 1 ∞ 1
𝑦 𝑦2 𝑦 𝑦3 𝑦2 𝑦3 𝑦2 𝑦4
1 1 1 1
= ∫ 𝑥 3 𝑓 (𝑥)𝑑𝑥 = ∫ 𝑥 3 . 𝑑𝑥 = [𝑥 4 ]1−1 = [1 − 1] = 0
= ∫3[ + − − ] 𝑑𝑦 = 3 [ + − − ] −∞ −1 2 2×4 8
3 2 4 2 6 6 8 8 0
0

1 1 1 1 1 1 4−3 1
= 3 [ + − − ] = 3 [ − ] = 3[ ]= 𝑐𝑜𝑣 (𝑋, 𝑌) 𝐸(𝑋𝑌 ) − 𝐸 (𝑋). 𝐸(𝑌) 0−0
6 6 8 8 3 4 12 4 𝑟𝑥𝑦 = = = = 0
√𝜎𝑥 . √𝜎𝑦 𝜎𝑥 . 𝜎𝑦 𝜎𝑥 . 𝜎𝑦
1 1 1
𝐶𝑜𝑣(𝑋, 𝑌) = 𝐸 (𝑋𝑌 ) − 𝐸(𝑋). 𝐸(𝑌) = − . = 0.
4 2 2
𝑐𝑜𝑣(𝑋,𝑌) 0
𝑟𝑥𝑦 = 𝜎𝑥.𝜎𝑦
= 𝜎𝑥 𝜎𝑦
= 0Hence X and Y are uncorrelated. Problems for Practice:

Since 𝑓 (𝑥 , 𝑦) = 𝑓(𝑥) . 𝑓 (𝑦) , X and Y are independent 1. The joint probability density function of X and Y is given below. Find
1 1 the correlation coefficient of X and Y.
Here 3 [(𝑥 + 𝑦) − (𝑥 2 + 𝑦 2 )] ≠ 9 [𝑥 − 𝑥 2 + ] [𝑦 − 𝑦 2 + ]
6 6
Y -1 1
Hence X and Y are not independent.
X

1 0 1 3
Problem: 5 Let X be a random variable with p.d.f𝑓(𝑥) = 2 , −1 ≤ 𝑥 ≤ 1 and
8 8
let 𝑌 = 𝑋 2. Find the correlation between X and Y 1 2 2
𝑐𝑜𝑣(𝑋,𝑌) 8 8
Solution: To find 𝑟𝑥𝑦 = 𝜎𝑥.𝜎𝑦
-------------------- (1)

First we find 𝐸(𝑋), 𝐸(𝑌), 𝐸(𝑋𝑌) 1 1 1

ANS: 𝐸(𝑋) = 2 ; 𝐸 (𝑌) = 4 ; 𝐸(𝑋 2 ) = 2 ; 𝐸(𝑌 2 ) = 1 ; 𝐸(𝑋𝑌) =
∞ 1 1
1 1 1 1 1 −1
𝐸(𝑋) = ∫ 𝑥𝑓(𝑥) 𝑑𝑥 = ∫ 𝑥 𝑑𝑥 = ∫ 𝑥 𝑑𝑥 = [𝑥 2 ]1−1 = [1 − 1] ; 𝑟𝑋𝑌 = √15 ;
2 2 2×2 4 2
−∞ −1 −1
2. Suppose that the two dimensional random variable (X,Y) has p.d.f
= 0
given by 𝑓(𝑥, 𝑦) = 2, 0 ≤ 𝑥 ≤ 𝑦 ≤ 1 𝑎𝑛𝑑 = 0, elsewhere. Find the
correlation co-efficient 𝑟𝑋𝑌 .
∞ 1
1 1 1 1 2 1 1
𝐸(𝑌) = 𝐸(𝑋 2 ) = ∫ 𝑥 2 𝑓(𝑥) 𝑑𝑥 = ∫ 𝑥 2 𝑑𝑥 = [𝑥 3 ]1−1 = [1 − (−1)] ANS: 𝐸(𝑋) = 3 ; 𝐸 (𝑌) = 3 ; 𝐸(𝑋 2 ) = 6 ; 𝐸(𝑌 2 ) = 2 ; 𝐸 (𝑋𝑌) =
2 2×3 6
−∞ −1 1 1 1 1 1
; 𝑉𝑎𝑟(𝑋 ) = ; 𝑉𝑎𝑟(𝑌) = ; 𝐶𝑜𝑣(𝑋, 𝑌) = ; 𝑟𝑋𝑌 = ;
= 1/3 4 18 18 36 2

𝐸(𝑋𝑌) = 𝐸(𝑋. 𝑋 2 ) = 𝐸[𝑋 3 ] 3. The independent random variables X and Y defined by

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 29
 4𝑎𝑥, 0 ≤ 𝑥 ≤ 𝑟 4𝑏𝑦, 0 ≤ 𝑦 ≤ 𝑠 r = ±√byx bxy , x̅ = meanof x ; y̅ = meanof y ; σx = S. D of x ; σy =
𝑓 (𝑥) = { 𝑓 (𝑦) = {
0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
S. D of y ;σx 2 = Var x ;σy 2 = Var y.
Find the correlation co-efficient between 𝑈 = 𝑋 + 𝑌 𝑎𝑛𝑑 𝑉 = 𝑋 − 𝑌.
1 1 2𝑟 𝑟2
NOTE: Regression coefficient are independent of change of origin but
ANS: 𝑎 = ;𝑏 = ; 𝐸 (𝑋 ) = ; 𝐸(𝑋 2 ) = ; 𝑉𝑎𝑟(𝑋) =
2𝑟 2 2𝑠 2 3 2 not of scale. i.e, 𝑏𝑦𝑥 = 𝑏𝑉𝑈 and 𝑏𝑥𝑦 = 𝑏𝑈𝑉
1 1 𝑏−𝑎
36𝑎
; 𝑉𝑎𝑟(𝑦) = 36𝑏 ; 𝑟(𝑈, 𝑉) = 𝑏+𝑎. 𝑛 ∑ 𝑈𝑉 − (∑ 𝑈)(∑ 𝑉) 𝑛 ∑ 𝑈𝑉 − (∑ 𝑈)(∑ 𝑉)
= =
REGRESSION: 𝑛 ∑ 𝑈 2 − (∑ 𝑈)2 𝑛 ∑ 𝑉 2 − (∑ 𝑉)2

Regression is a mathematical measure of the average relationship between Where U = X − a , V = Y − b where a, b are arbitrary values.

two or more variables in terms of the original limits of the data.

Lines of regression [Formula]: PROPERTIES OF REGRESSION LINE:

1. The line of regression of Y on X is given by 1. Correlation coefficient is the geometric mean between regression
𝜎𝑦 coefficients.
𝑦 − 𝑦̅ = 𝑟 (𝑥 − 𝑥̅ )
𝜎𝑥 i.e, 𝑟 = √𝑏𝑥𝑦 × 𝑏𝑦𝑥
2. The line of regression of X and Y is given by
2. The correlation coefficient and the two regression coefficients have same
𝜎𝑥
𝑥 − 𝑥̅ = 𝑟 (𝑦 − 𝑦̅) sign
𝜎𝑦
3. The regression lines pass through the point (𝑥̅ , 𝑦̅) ,
Note:
σy Where 𝐱̅ = mean of X and 𝐲̅ = mean of Y.
1. The quantity r is called regression coefficient of Y on X and is
σx
Problem: 1
denoted by 𝑏𝑦𝑥 Obtain the equation of the lines of regression for the following data:
σy
𝑖. 𝑒., 𝑏𝑦𝑥 =r X 65 66 67 67 68 69 70 72
σx
σ Y 67 68 65 68 72 72 69 71
2. The quantity r σx is called regression coefficient of X on Y and is
y
Solution:
denoted by 𝑏𝑥𝑦 Let𝑢 = 𝑋 – 69:&𝑣 = 𝑌 – 69;
σx
𝑖. 𝑒., 𝑏𝑥𝑦 = r First ,we find regression coefficients
σy
𝑛 ∑ 𝑈𝑉 − (∑ 𝑈 )(∑ 𝑉)
Correlation coefficient: 𝑏𝑥𝑦 = 𝑏𝑢𝑣 = ℎ𝑒𝑟𝑒 𝑛 = 8
𝑛 ∑ 𝑉 2 − (∑ 𝑉)2

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 30
 8 × 24 − (−8)(0) 8 × 24 Also the coefficient of correlation between the two is 0.8. Estimate the
= 2
= = 0.5454
8 × 44 − (0) 8 × 44 most likely price of rice (i) at Chennai corresponding to the price of 18 at
𝑛 ∑ 𝑈𝑉 − (∑ 𝑈)(∑ 𝑉) Madurai & (ii)at Madurai corresponding to the price of 17 at Chennai.
𝑏𝑦𝑥 = 𝑏𝑣𝑢 =
𝑛 ∑ 𝑈 2 − (∑ 𝑈)2
Solution:
8 × 24 − (−8)(0) 8 × 24 8 × 24
= 2
= = = 0.667 Let the prices of rice at Chennai and Madurai be denoted by X and Y
8 × 44 − (−8) 352 − 64 288
∑ 𝑋 544 ∑ 𝑌 552 respectively.
𝑥̅ = = = 68 𝑎𝑛𝑑 𝑦̅ = = = 69 From the given data,
𝑛 8 𝑛 8
Equation of the line of regression of Y on X is 𝑥 ̅ = 19.5, 𝑦 ̅ = 17.75, 𝜎𝑥 = 1.75, 𝜎𝑦 = 2.5 𝑎𝑛𝑑 𝑟 = 0.8
𝑦 − 𝑦̅ = 𝑏𝑦𝑥 (𝑥 − 𝑥̅ ) Regression line of X on Y is
𝑦 – 69 = 0.667(𝑥 – 68) 𝑥 − 𝑥̅ = 𝑏𝑥𝑦 (𝑦 − 𝑦̅)
𝑦 = 0.667𝑥 – (0.667)(68) + 69 𝜎𝑥
𝑥 − 𝑥̅ = 𝑟 (𝑦 − 𝑦̅)
𝜎𝑦
𝑦 = 0.667𝑥 + 23.64
1.75
Equation of the line of regression of X on Y is 𝑥 − 19.5 = (0.8) (𝑦 – 17.75)
2.5
𝑥 − 𝑥̅ = 𝑏𝑥𝑦 (𝑦 − 𝑦̅)
When 𝑦 = 18 , then
𝑥 – 68 = 0.545(𝑦 – 69) 0.8 × 1.75
𝑥 – 19.5 = × 0.25
𝑥 – 68 = 0.545𝑦 – (0.545)(69) 2.5
𝑥 = 0.545𝑦 + 30.395 0.8 × 1.75 × 0.25
𝑥 = 19.5 +
2.5
Problem: 2
𝑥 = 19.64
A Study of prices of rice at Chennai and Madurai gave the following
Regression line of Y on X is
data:

Chennai Madurai
Mean 19.5 17.75
S.D 1.75 2.5
𝑦 − 𝑦̅ = 𝑏𝑦𝑥 (𝑥 − 𝑥̅ )
𝜎𝑦
𝑦 − 𝑦̅ = 𝑟 (𝑥 − 𝑥̅ )
𝜎𝑥
Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 31
 (0.8)(2.5)(𝑥 − 19.5) 8 66 18 214
𝑦 – 17.75 = 𝑦= 𝑥+ 𝑥 = 𝑦+
1.75 10 10 40 40
8 18
When 𝑋 = 17 , then i.e, 𝑏𝑦𝑥 = 𝑎𝑛𝑑 𝑏𝑥𝑦 =
10 40
0.8 × 2.5 × −2.5 2
𝑦 = 17.75 + we know that 𝑟 = 𝑏𝑦𝑥 × 𝑏𝑥𝑦
1.75
8 18
𝑦 = 14.89 𝑟2 = × = 0.36
10 40
Problem: 3
𝑟 = ± 0.6
The two line of regression are 8𝑥 – 10𝑦 + 66 = 0, 40𝑥 – 18𝑦 – 214 =
Since, both the regression lines are positive, correlation coefficient
0.The variance of x is 9. Find (i) The mean values of x and y (ii)Correlation
r(X,Y) must be positive.
coefficient between x and y. (iii) Standard deviation of y.
 𝑟 = 0.6
Solution:
(iii) Given var (x) = 9
(i) Since both the lines of regression passes through x̅ and y̅, the point
σx 2 = 9 σx = 3
must satisfy the two regression lines. 𝜎𝑦
We know that 𝑏𝑦𝑥 = 𝑟 𝜎
Now , 8𝑥̅ − 10𝑦̅ = −66 − − − −(1) 𝑥

8 𝜎𝑦
40𝑥̅ − 18𝑦̅ = 214 − − − −(2) = 0.6 ( )
10 3
(1)×540𝑥 ̅ − 50𝑦 ̅ = −330 3×8 8
𝜎𝑦 = = =4
0𝑥 ̅ − 18𝑦 ̅ = 214 10 × 0.6 2
−32𝑦 ̅ = −544 Standard deviation of y isσy = 4.
𝑦 ̅ = 544/32 = 17 Problem: 4
𝑃𝑢𝑡 𝑦̅ = 17in eqn. (1) , we get 8𝑥̅ − 170 = −66 Two regression lines are 4𝑥 − 5𝑦 + 33 = 0 𝑎𝑛𝑑 20𝑥 − 9𝑦 = 107 and
104 variance of x is 25 . Find the means of x and y. Also, find the coefficient
8𝑥̅ = −66 + 170 = 104𝑥̅ = = 13
8 correlation.
Hence , the mean values of x and y are 𝑥 ̅ = 13 ; 𝑦 ̅ = 17
Solution:
(ii) Let us suppose that 8𝑥 − 10𝑦 = −66 & 40𝑥 − 18𝑦 = 214 be the
Since both the lines of regression passing through (𝑥̅ , 𝑦̅ ) ,
line of regression Y on X and X on Y respectively .
We get, 4𝑥̅ − 5𝑦̅ = −33 − − − − − (1)
8𝑥 – 10𝑦 = −66 &40𝑥 − 18𝑦 = 214
20𝑥̅ − 9𝑦̅ = 107 − − − −(2)
10𝑦 = 8𝑥 + 66 40𝑥 = 18𝑦 + 214
Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 32
(1)×5 20𝑥̅ − 25𝑦̅ = −165 5𝑥 ̅ = 132 + 108 = 240 ⟹ 𝑥 ̅ = 48
20𝑥̅ – 9𝑦̅ = 107 Let 3𝑦 − 5𝑥 = −108 be the line of regression of X and Y respectively.
−16𝑦̅ = −272 i.e., 5𝑥 = 3𝑦 + 108
𝑦̅ = 17 ⟹ 𝑥 = (3/5)𝑦 + (108/5) 𝑏𝑥𝑦 = 3/5
𝑝𝑢𝑡 𝑦̅ = 17 𝑖𝑛 𝑒𝑞𝑛. (1), 𝑤𝑒𝑔𝑒𝑡 4𝑥̅ − 5(17) = −33 𝑟(𝜎𝑥 / 𝜎𝑦 ) = 3/5 − − − −(1) ∵ 𝑏𝑥𝑦 = 𝑟(𝜎𝑥 /𝜎𝑦 )
4𝑥̅ = −33 + 85 Given that the variance of 𝑋 =
9
th of the variance of Y
16
4𝑥̅ = 52
𝜎𝑥2 = (9/16). 𝜎𝑦2 ⇒ 𝜎𝑥 = (3/4). 𝜎𝑦 − − − −(2)
𝑥̅ = 13
sub(2)in (1), weget
Therefore the mean values are𝑥̅ = 13 ; 𝑦̅ = 17
𝑟 3 3 3 4 4
(i) Let 4𝑥 − 5𝑦 = −33 & 20𝑥 − 9𝑦 = 107 be the line of ( ). ( ) . 𝜎𝑦 = ⟹ 𝒓 = ( )( ) = ⟹ 𝑟 = 0.8
𝜎𝑦 4 5 5 3 5
regression of Y on X and X on Y respectively.
Problem: 6
5𝑦 = 4𝑥 + 33 ; 20𝑥 = 9𝑦 + 107
Can 𝑦 = 5 + 2.8𝑥 and 𝑥 = 3 − 0.5𝑦 be the estimated regression equation of
4𝑥 33 9𝑦 107
𝑦 = 5
+ 5
; 𝑥 = 20
+ 20 Y on X and X on Y respectively? Explain your answer.
4 9
𝑏𝑦𝑥 = ; 𝑏𝑥𝑦 = Solution:
5 20
4 9 9 3 Given 𝑥 = 3 − 0.5𝑦
We know that 𝑟 2 = 𝑏𝑦𝑥 × 𝑏𝑥𝑦 = 5
× 20 = 25
= 0.36 𝑟 = 5
= 0.6
𝑏𝑥𝑦 = −0.5
Problem: 5
𝑦 = 5 + 2.8𝑥
The regression equation of X and Y is 3y − 5x + 108 = 0. If the mean value
th
𝑏𝑦𝑥 = 2.8
9
of y is 44 & the variance of X were (16) of the variance of Y. Find the mean
we know that 𝑟 2 = 𝑏𝑥𝑦 × 𝑏𝑦𝑥 = −0.5 × 2.8 = −1.4
value of X & the correlation coefficient. 𝑟 = √−1.4which is imaginary
Solution: 𝑟 cannot be imaginary
Since the line regression passing through (𝑥̅ , 𝑦), we get
∴The given lines are not estimated as regression equations.
3¯𝑦 − 5𝑥 ̅ = −108
3(44) − 5𝑥 ̅ = −108
132 − 5𝑥 ̅ = −108

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 33
Problems for Practice: Are X and Y independent? Find 𝑃 (1 < 𝑋 < 3 , 1 < 𝑌 < 2).
1. Obtain the equations of the lines of regression from the following 2. Find the coefficient of correlation between X and Y from the data
data: given below:
X 1 2 3 4 5 6 7 X 65 66 67 67 68 69 70 72
Y 9 8 10 12 11 13 14 Y 67 68 65 68 72 72 69 71
3. The two line of regression are 8𝑋 – 10𝑌 + 66 = 0,
2. The regression equations of X and Y are respectively 5𝑥 − 𝑦 = 22 and 40𝑋 – 18𝑌 – 214 = 0. The variance of X is 9. Findthe mean values
64𝑥 − 45𝑦 = 24. Find the means of X and Y. of X and Y. Also find the coefficient of Correlation between the
Ans:𝑥̅ = 6 , 𝑦̅ = 8 variablesX and Y.
3. Given the two regression lines 3𝑥 + 12𝑦 = 19, 3𝑦 + 9𝑥 = 46, find the 4. Two random variables X and Y have the following joint probability
coefficient of correlation between X and Y. 𝑥 + 𝑦 ∶ 0 ≤ 𝑥 ≤ 1, 0 ≤ 𝑦 ≤ 1
density function. 𝑓 (𝑥, 𝑦) = { . Find the
0 ∶ 𝑂𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
−1 −1 1
Ans:𝑏𝑦𝑥 = 4
, 𝑏𝑥𝑦 = 3
, 𝑟 = −√12 probability density function of the random variable 𝑈 = 𝑋𝑌.
Dec – 2014
PART – A
SAMPLE UNIVERSITY QUESTIONS
1. The joint p.d.f of two dimensional random variables (X , Y) is given by
MAY-2015
𝑓 (𝑥, 𝑦) = 𝐾𝑥𝑒 −𝑦 ; 0 < 𝑥 < 2 , 𝑦 > 0. Find the value of K
PART – A
2 +𝑦 2 )
2. Comment on the statement: “ If𝐶𝑂𝑉 ( 𝑋 , 𝑌 ) = 0 then X and Y are
1. Find the value of k, if 𝑓(𝑥, 𝑦) = 𝑘𝑥𝑦𝑒 −(𝑥 : 𝑥 ≥ 0, 𝑦 ≥ 0 is to be
uncorrelated “.
joint probability density function.
PART – B
2. What is the angle between the two regression lines?
1. Calculate the coefficient of correlation for the following data:
PART – B
X : 9 8 7 6 5 4 3 2 1
1. If the joint distribution function of two dimensional random
Y : 15 16 14 13 11 12 10 8 9
variable (X,Y) is given by 𝐹(𝑥, 𝑦) =
(1 − 𝑒 −𝑥 )(1 − 𝑒 −𝑦 ) 𝑓𝑜𝑟 𝑥 > 0, 𝑦 > 0
{ Find the marginal densities of X
0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
and Y.
Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 34
 1. The joint p.d.f X and Y is given by 𝑓(𝑥, 𝑦) =
May – 2014 1
(6 − 𝑥 − 𝑦), 0 < 𝑥 < 2,2 < 𝑦 < 4
{8 Find(𝑖 ) 𝑃(𝑋 < 1 ∩ 𝑌 < 3)(𝑖𝑖 ) 𝑃 (𝑋 <
PART – A 0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
1. Find the value of k, if the joint density function of (X , Y) is given by 1/𝑌 < 3), (𝑖𝑖𝑖) 𝑃(𝑋 + 𝑌 < 3)
𝑘 (1 − 𝑥)(1 − 𝑦), 0 < 𝑥 < 4 , 1<𝑦<5 2. If X and Y each follow an exponential distribution with
𝑓 (𝑋 , 𝑌) = { }
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Parameter 1 and are independent, find the p.d.f of 𝑈 = 𝑋 – 𝑌 . The
2. Given that joint probability density function of (𝑋 , 𝑌) as
marks obtained by 10 students in Mathematics and Statistics are given
1
𝑓 (𝑋 , 𝑌) = , 0 < 𝑥 < 2, 0 < 𝑦 < 3, Determine the marginal density. below.
6

PART – B Find the correlation coefficient between the two subjects.

1. The joint probability density of two random variables X and Y is given Maths 75 30 60 80 53 35 15 40 38 48
6
[𝑥 2 +
𝑥𝑦
] , 0 ≤ 𝑥 ≤ 1, 0 ≤ 𝑦 ≤ 2 Statistics 85 45 54 91 58 63 35 43 45 44
by 𝑓 (𝑥 , 𝑦) = {7 2 . Find the
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 3. A distribution with unknown mean μ has variance equal to 1.5. Use
conditional density function of X given Y and the conditional density central limit theorem to find how large a sample should be taken from
function of Y given X. the distribution in order that the probability will be atleast 0.95 that the
2. If the independent random variables X and Y have the variances 36 and sample mean will be within 0.5 of the population mean.
16 respectively , find the correlation coefficient rUV where U = X + Y and May – 2013
V = X – Y. PART – A
3. The joint probability density function of two random variables X and Y 1. If the joint p.d.f of (𝑋, 𝑌) is given by 𝑓 (𝑥, 𝑦) = 2, 𝑖𝑛 0 ≤ x < 𝑦 ≤ 1.
𝑘[(𝑥 + 𝑦) − (𝑥 2 + 𝑦 2 )], 0 < (𝑥 , 𝑦) < 1}. Find 𝐸(𝑋).
is given by 𝑓 (𝑥 , 𝑦) = { Show that
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 PART – B
X and Y are uncorrelated but not independent. 3
(𝑥 2 + 𝑦 2 ), 0 ≤ 𝑥 ≤ 1 , 0 ≤ 𝑦 ≤ 1
Dec – 2013 1. If 𝑓(𝑥, 𝑦) = {2 , Find the correlation
0 , 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
PART – A coefficient betweenγxy .
−(𝑥 2+𝑦 2 )
1. The joint p.d.f of R.V X and Y is given by 𝑓(𝑥, 𝑦) = 𝑘𝑥𝑦𝑒 ,𝑥 > 𝑥+𝑦
2. The joint distribution of X and Y is given by (𝑥 , 𝑦) = 21
, 𝑥 = 1,2,3; 𝑦 =
0, 𝑦 > 0. Find the value of 𝑘.
1,2 . Find the marginal distributions.
PART – B

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 35
3. If the p. d. f of X is 𝑓𝑋 (𝑥) = 2𝑥, 0 < 𝑥 < 1, find the pdf of 𝑌 = 3𝑋 + 1. PART – B
1. Let X and Y be two random variables having the joint probability
Dec – 2012 function given by 𝑓(𝑥 , 𝑦) = 𝑘(2𝑥 + 3𝑦), where X & Y can assume only the
PART – A integer values 0, 1 & 2. Find the marginal and conditional distributions
1. When will the two regression lines be (a) at right angle (b) coincident? 2. The two random variables X and Y have the joint probability density
2. A small college has 90 male and 30 female professors. An ad-hoc 𝐶 (4 − 𝑥 − 𝑦), 0 ≤ 𝑥 ≤ 2 , 0 ≤ 𝑦 ≤ 2
function 𝑓(𝑥, 𝑦) = { Find 𝐶𝑜𝑣(𝑋, 𝑌).
0 , 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
committee of 5 is selected at random o unity the vision and mission of the
3. Let X and Y be two random variables with the joint p.d.f𝑓 (𝑥, 𝑦) =
college. If X and Y are the number of men and women in the committee
8𝑥𝑦, 0 < 𝑥 < 𝑦 < 1 1 1
respectively what is the probability mass function of X and Y? { (i) Find 𝑃(𝑋 < /𝑌 < )(ii) Find the marginal and
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 2 4

conditional distributions.
PART – B (iii)Verify whether X and Y are independent.
1. Obtain the equation of the lines of regression from the following data Dec – 2011
X 1 2 3 4 5 6 7 PART – A
2 +𝑦 2 )
Y 9 8 10 12 11 13 14 1. The joint p.d.f of R.V X and Y is given by 𝑓(𝑥, 𝑦) = 𝑘𝑥𝑦𝑒 −(𝑥 ,𝑥 >
0, 𝑦 > 0. Find the value of 𝑘.
2. The joint p.d.f of a bivariate R.V (X,Y) is given by
𝜆𝑥𝑦, 0 ≤ 𝑥 ≤ 𝑦 ≤ 1 2. Given the random variable X with density function 𝑓 (𝑥) =
𝑓(𝑥, 𝑦) = { (1) Determine the value of λ . (2) Find the
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 2𝑥, 0 < 𝑥 < 1
{ Find the p.d.f of 𝑌 = 8𝑋 3 .
0, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
marginal and conditional probability density function of X.
PART – B
3. The regression equation of x and y is 3𝑦 − 5𝑥 + 108 = 0. If the mean
1. If the joint distribution of x and y is given by 𝑝(𝑥 , 𝑦) =
9 𝑡ℎ
value of y is 44 and the variance of x were( ) of the variance of y. Find the
16 𝑘 (2𝑥 + 3𝑦), 𝑦 = 1,2,3 ; 𝑥 = 0,1,2.Find (i) all the marginal distributions, (ii)
mean value of x and the correlation coefficient. all the conditional distribution
May – 2012 2. If X and Y are independent random variables with p.d.f’s 𝑒 −𝑥 , 𝑥>
PART – A 0 𝑎𝑛𝑑 𝑒 −𝑦 , 𝑦 > 0 respectively, find the density function of 𝑈 =
𝑋
and
𝑋+𝑌
1. The regression equations of X and Y are respectively 5𝑋 – 𝑌 = 22
𝑉 = 𝑋 + 𝑌 . Are U and V are independent?
and 64𝑋 – 45𝑌 = 24. Find the means of X and Y.

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 36
3. Find the coefficient of correlation between X and Y using the following Dec – 2010
data PART – A
2 +𝑦 2 )
1. Let the joint p.d.f of R.V X and Y is given by f(𝑥, 𝑦) = 4𝑥𝑦𝑒 −(𝑥 ,𝑥 >
X 10 14 18 22 26 30 0, 𝑦 > 0. Are X and Y are independent? Why or why not?
Y 18 12 24 63 30 36

PART – B
1. Compute the coefficient of correlation between X and Y using the data
May – 2011 X 1 3 5 7 8 10
PART – A Y 8 12 15 17 18 20
1. Given the two regression lines3𝑋 + 12𝑌 = 19, 3𝑌 + 9𝑋 = 46, find the
coefficient of correlation between X and Y. May – 2010
PART – B PART – A
1. Given the joint density function 𝑓(𝑥, 𝑦) = 1. Give a real life example each for positive correlation and negative
(1+3𝑦 2 )
correlation
{𝑥 4
, 0 < 𝑥 < 2, 0 < 𝑦 < 1. Find the marginal densities 𝑔(𝑥), ℎ(𝑦)
0, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒 PART – B
1 1 1
and the conditional density 𝑓(𝑥⁄𝑦) and evaluate 𝑃[4 < 𝑥 < 2⁄𝑦 = 3] 1. The joint probability of X and Y is given by𝑓𝑋𝑌 (𝑥, 𝑦) =
1
2. Determine whether the random variables X and Y are independent, given 𝑦𝑒 −𝑥𝑦 , 0 < 𝑥 < ∞, 0 < 𝑦 < 2
{2 . Calculate the conditional density of X
their joint probability density function as 𝑓 (𝑥, 𝑦) = 0 , 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑥𝑦 given 𝑌 = 1.
𝑥2 + , 0 ≤ 𝑥 ≤ 1, 0 ≤ 𝑦 ≤ 2
{ 3
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 2. If the correlation coefficient is 0, then the can we conclude that they
3. If X and Y are independent random variables of a density functions are independent? Justify your answer, through an example. What about

2𝑒 −2𝑥 , 𝑥 ≥ 0 3𝑒 −3𝑦 , 𝑦 ≥ 0 the converse?

𝑓 (𝑥) = { and 𝑓 (𝑦) = { respectively, find the
0, 𝑥<0 0, 𝑦<0 3. Let X and Y be independent random variables both uniformly
density functions of𝑍 = 𝑋 – 𝑌. Distributed on (0, 1) Calculate the probability density of 𝑋 + 𝑌.

Dr.A. Manickam, Associate Professor, School of Sciences, Division of Mathematics, FET, SRMIST, Trichy -621105, Tamilnadu, India. Page 37
 CENTRAL LIMIT THEOREM EXAMPLES:
1.The life of a certain brand of a tube light may be considered as a
LIAPOUNOFF’S FORM OF CENTRAL LIMIT THEOREM Random Variable with mean 1200h and standard deviation 250h. Find
If 𝑥1 , 𝑥2 , ………….. 𝑥𝑛 be a sequence of independent R.V’s with the probability. Using central limit theorem, that the average life time
E(𝑋𝑖 )= 𝜇𝑖 and Var (𝑋𝑖 ) = 𝜎𝑖2 , 𝑖 = 1,2, … … 𝑛 and if 𝑆𝑛 = 𝑥1 , 𝑥2 , of 60 lights exceeds 1250h.
…………..𝑥𝑛 , then under certain general conditions, Sn follows a normal Solution:
distribution with mean ∑𝑛𝑖=1 𝜇𝑖 and variance 𝑎2 = ∑𝑛𝑖=1 𝜎𝑖2 as n → ∞. Given that Mean 𝜇= 1200 h, S.D. 𝜎= 250 h ; n = 60 and 𝑋̅ =mean life
time of 60 lights.
LINDEBERG – LEVY’S FORM OF CENTRAL LIMIT THEOREM By central limit theorem, we have
If 𝑥1 , 𝑥2 , ………….. 𝑥𝑛 be a sequence of independent identically 𝜎 250
𝑋̅~ 𝑁 (𝜇, ) ~ 𝑁 (1200, )
√𝑛 √60
distributed RV’s with E(𝑋𝑖 ) = 𝜇 and Var (𝑋𝑖 ) = 𝜎 2 , 𝑖 =
To find P (𝑋̅ > 1250)
1,2, … … 𝑛 𝑎𝑛𝑑 𝑖𝑓 𝑆𝑛 =.𝑥1 + 𝑥2 + ⋯ + 𝑥𝑛 then under certain general 𝑋̅− 𝜇 𝑋̅− 1200
conditions, ‘Sn’ follows a normal distribution with mean n𝜇 and variance Let z = 𝜎 = 250
√𝑛 √60
n𝜎 2 as n → ∞.
𝑋̅− 1200 1250 − 1200
COROLLARY FORM OF CENTRAL LIMIT THEOREM Now, P (𝑋̅ > 1250) =( 250 > 250 )
√60 √60
1
If 𝑋̅ = 𝑛 = (𝑋1 + 𝑋1 . . . . . . 𝑋𝑛 ), then E (𝑋̅) = 𝜇 and Var (𝑋̅) = 50
= 𝑃(𝑧 > 250 )
1 𝜎2
( 2) =
2 𝑛𝜎
√60
𝑛 𝑛
𝜎2 = P ( z > 1.55)
∴ 𝑋̅ follows a normal distribution with mean µ and variance as n → ∞.
𝑛 P (𝑋̅ > 1250) = 0.0606.
Remark ( Formula ) 2.A Random sample of size 100 is taken from a population whose mean
(1) If 𝑋̅ is given then we use the following formula is 60 and variance is 400. Using central limit theorem, with what
𝜎 𝑋̅−𝜇
𝑆𝑛 ~ 𝑁 (𝜇, ) and z = 𝜎 probability can we assert that the mean of the sample will not differ
√𝑛
√𝑛
from = 60 by more than 4.
(2) If Sn is given then we use
𝑆𝑛−𝑛𝜇
Solution:
Sn ~ 𝑁 (𝑛𝜇, 𝜎√𝑛) and z = Given Mean 𝜇 = 60, Variance 𝜎 2 = 400 ; n=100 S.D., 𝜎=20
𝜎 √𝑛
By central limit theorem,
𝜎
𝑋̅ ~ N (𝜇 )
√𝑛

Dr.A. Manickam School of Advanced Science and Languages. VIT Bhopal University ,Kothari kalan-466114,Sehore (Dt),Madhya Pradesh,India Page 1
 20 ∵ 𝑆𝑛 = 𝑥1+ 𝑥2 + ⋯ +𝑥75 .
~ N (60 )
√100
20 ~ 𝑁 (75 𝑋 2, √2√75)
~ N (60 )
√10 ~ 𝑁 (150, √150)
~ N (60, 2) 𝑆𝑛−𝑛𝜇 𝑆𝑛−150
̅−𝜇
𝑋 ̅ − 60
𝑋
Let z = =
𝜎 √𝑛 √150
Let z = 𝜎 = 2 To find P ( 120 < Sn < 160 )
√𝑛

To find P (|𝑋̅ − 𝜇| ≤ 4) P ( 120 < Sn < 160 ) = 𝑃 (

120−150
<
𝑆𝑛−150
<
160−150
)
√150 √150 √150
i.e., P ( Mean of the sample will not differ from 𝜇 = 60 by more −30 10
than 4 ) = 𝑃 (|𝑋̅ − 𝜇| ≤ 4) = 𝑃 <𝑧<
√150 √150
= 𝑃 (−4 ≤ 𝑋̅ – 𝜇 ≤ 4 ) ∵ = 𝑃 ( −2.45 < 𝑧 < 0.85)
|𝑋 | ≤ 1 = 𝑃 (−2.45 < 𝑧 < 0.85) + 𝑃 (0 < 𝑧 < 0.85)
4 ̅ – 60
𝑋
= 𝑃 (− 2 ≤ ≤ 4) -1 = 𝑃 (0 < 𝑧 < 2.45) + 𝑃 (0 < 𝑧 < 0.85)
2
|≤ 𝑋 ≤| = 0.4929 + 0.3023 = 0.7952.
̅ − 60
𝑋 4.If 𝑿𝒊 , i = 1,2, . . . . . 50 are independent R.V’s each having a poisson
= 𝑃 (−2 ≤ 𝑧 ≤ 2) ∵ 𝑧 = 2
distribution with parameter𝝀 = 𝟎. 𝟎𝟑 and Sn =𝑿𝟏 + 𝑿𝟐 + … … + 𝑿𝒏
= 2𝑃 ( 0 ≤ 𝑧 ≤ 2 )
evaluate P ( Sn ≥ 3 )
= 2 ( 0.4773 )
Solution:
= 0.9546.
For Poisson distribution, Mean = Variance = λ. Here λ = 0.03
3.If 𝒙𝟏 , 𝒙𝟐 , ………….. 𝒙𝒏 are poisson variables with parameter x = 2, use
From this, we get
the central limit the theorem to estimate P ( 120 < 𝑺𝒏 < 160 ) Where 𝑺𝒏
Mean µ = 0.03, Variance 𝜎 2 =0.03 & n = 50
= 𝒙𝟏 + 𝒙𝟐 + ⋯ + 𝒙𝒏 and 𝒏 = 𝟓
Solution: ⇒ 𝜎 = √0.03
Given that 𝑥1 , 𝑥2 , …………..𝑥𝑛 are poisson Variables for poisson By central limit theorem,
distribution. Sn ~ (𝑛𝜇, 𝜎√𝑛 )
Mean = Veriance =λ , here λ = 2. ~ 𝑁(50 𝑋 0.03, √0.03√50 )
From this, we get ~ 𝑁(1.5, √1.5 )
𝑆𝑛−𝑛𝜇 𝑆𝑛−1.5
Mean = 2 ; Variance 𝜎 2 = 2 ; n =75 Let z = =
𝜎 √𝑛 √1.5
⇒ 𝜎 = √2 To find P ( Sn ≥ 3 )
By Central limit theorem, 𝑆𝑛−1.5 3 −1.5
P ( Sn ≥ 3 ) = P ( ≥ )
Sn ~ 𝑁 (𝑛𝜇, 𝜎 √𝑛) √1.5 √1.5

Dr.A. Manickam School of Advanced Science and Languages. VIT Bhopal University ,Kothari kalan-466114,Sehore (Dt),Madhya Pradesh,India Page 2
 = P ( z ≥ 1.22 )
= 0.1112.
5. Suppose that in a certain circuit, 20 resistors are connected in series.
The mean and variance of each resistor are 5 and 0.20 respectively.
What is the probability that the total resistance of the circuit will exceed
98 ohms, assuming independence.
Solution:
Let Sn =𝑋1 + 𝑋2 + … … + 𝑋20 Since, the resistors are connected
in series.
Given that Mean = 5 ; Variance 𝜎 2 = 0.20 & n = 20
⇒ 𝜎 = √0.20
By central limit theorem,
Sn ~ 𝑁 (𝑛𝜇, 𝜎√𝑛 )
~ 𝑁(20 𝑋 5, (√0.20)(√20)
~ 𝑁 (100, √4 )
~ 𝑁 (100, 2 )
𝑆𝑛−𝑚𝜇
Let 𝑧 = 𝜎 √𝑛
𝑆𝑛 − 100
=
2
To find P ( Sn > 98 )
𝑆𝑛−100 98−100
Now, P ( Sn > 98 ) = P ( > )
2 2
= P ( z > −1 )
= 0.8413.

Dr.A. Manickam School of Advanced Science and Languages. VIT Bhopal University ,Kothari kalan-466114,Sehore (Dt),Madhya Pradesh,India Page 3