21-01-2026
Transformations of Multiple Random
Variables
Consider a random variable 𝒁 = 𝒈(𝑿, 𝒀) which is a function of two
random variables X and Y.
The expectation of g(X, Y) can be obtained using
In the above equation, to find expectation of g(X,Y), we used the
joint pdf of X and Y. That is, to find expectation of g(X,Y), it is not
necessary to find the density function of the new random variable Z.
However, in some practical problems, it may be required to
determine the density function of the transformed variable Z, given
the joint pdf of X and Y.
86 Dr. [Link], Associate Professor, SENSE, VIT
Two Functions of Two Random Variables
Consider two random variables X and Y with joint pdf 𝑓𝑋𝑌 (𝑥, 𝑦).
Let us define two random variables Z and W which arise as
functions of X and Y.
That is, 𝒁 = 𝒈(𝑿, 𝒀) and 𝑾 = 𝒉(𝑿, 𝒀).
Let us assume that the functions Z and W satisfy the following
conditions.
1. The equations 𝒁 = 𝒈(𝒙, 𝒚) and 𝑾 = 𝒉(𝒙, 𝒚) can be uniquely
solved for x and y in terms of z and w.
The solutions are given by 𝑿 = 𝒈(𝒛, 𝒘) and 𝒀 = 𝒉(𝒛, 𝒘)
87 Dr. [Link], Associate Professor, SENSE, VIT
1
21-01-2026
Two Functions of Two Random Variables
2. The functions g and h have continuous partial derivatives at all
points (x, y)
3. The determinant Covariance Matrix
Or
Jacobian Matrix
If the above conditions are satisfied then the random variables Z
and W are jointly continuous with density function
88 Dr. [Link], Associate Professor, SENSE, VIT
Tutorial Example
Let X and Y be jointly continuous random variables with pdf
𝑓𝑋𝑌 (𝑥, 𝑦).
Let 𝑍 = 𝑋 + 2𝑌 and 𝑊 = 𝑋– 2𝑌.
Find the joint pdf of Z and W in terms of 𝑓𝑋𝑌
SOLUTION
Let 𝑔(𝑋, 𝑌) = 𝑋 + 2𝑌
and ℎ(𝑋, 𝑌) = 𝑋– 2𝑌
89 Dr. [Link], Associate Professor, SENSE, VIT
2
21-01-2026
Tutorial Example
90 Dr. [Link], Associate Professor, SENSE, VIT
Auxiliary Variables
Consider the transformation 𝑍 = 𝑔(𝑋, 𝑌) where X and Y are two
random variables.
To determine the pdf of Z, first we define an auxiliary variable 𝑾 =
𝑿 𝒐𝒓 𝑾 = 𝒀 and
Then obtain the joint density of Z and W using
Next, we can obtain 𝑓𝑍 𝑧 from 𝑓𝑍𝑊 𝑧, 𝑤 by proper integration
with respect to ‘𝒘’.
91 Dr. [Link], Associate Professor, SENSE, VIT
3
21-01-2026
Tutorial Example
The random variables X and Y the have joint density function
Find the density function of 𝒁 = 𝑿 + 𝟐𝒀.
92 Dr. [Link], Associate Professor, SENSE, VIT
Tutorial Example
93 Dr. [Link], Associate Professor, SENSE, VIT
4
21-01-2026
Tutorial Example
94 Dr. [Link], Associate Professor, SENSE, VIT
Tutorial Example
95 Dr. [Link], Associate Professor, SENSE, VIT
5
21-01-2026
Tutorial Example
96 Dr. [Link], Associate Professor, SENSE, VIT
Tutorial Example
97 Dr. [Link], Associate Professor, SENSE, VIT
6
21-01-2026
Tutorial Example
98 Dr. [Link], Associate Professor, SENSE, VIT
Tutorial Example
99 Dr. [Link], Associate Professor, SENSE, VIT
7
21-01-2026
Tutorial Example
If X and Y are Independent random variables with density functions
𝑓𝑋 𝑥 = 𝑒 −𝑥 𝑢(𝑥); 𝑓𝑌 𝑦 = 2𝑒 −𝑦 𝑢 𝑦 . Find the density function of Z
= X + Y.
SOLUTION :Since X and Y are Independent
100 Dr. [Link], Associate Professor, SENSE, VIT
Tutorial Example
101 Dr. [Link], Associate Professor, SENSE, VIT
8
21-01-2026
Tutorial Example
The range space of (Z, W) can be obtained as follows:
102 Dr. [Link], Associate Professor, SENSE, VIT
Linear Transformation of Gaussian Random
Variables
Consider a set of N Gaussian random variables 𝑿𝟏 , 𝑿𝟐 , … , 𝑿𝑵 with
joint density function
Let these random variables be linearly transformed to a new set of
random variables 𝒀𝟏 , 𝒀𝟐 , … , 𝒀𝑵 using the following relation
103 Dr. [Link], Associate Professor, SENSE, VIT
9
21-01-2026
Linear Transformation of Gaussian Random
Variables
The above equation can be represented in matrix form as
104 Dr. [Link], Associate Professor, SENSE, VIT
Linear Transformation of Gaussian Random
Variables
whose inverse is given by
105 Dr. [Link], Associate Professor, SENSE, VIT
10
21-01-2026
Linear Transformation of Gaussian Random
Variables
From the above equation we can find
In general, we can write
106 Dr. [Link], Associate Professor, SENSE, VIT
Linear Transformation of Gaussian Random
Variables
We can write
107 Dr. [Link], Associate Professor, SENSE, VIT
11
21-01-2026
Linear Transformation of Gaussian Random
Variables
If [CX] represents the covariance matrix of random variables X1, X2,
…, XN then ijth element of the covariance matrix CXiXj can be
obtained as
108 Dr. [Link], Associate Professor, SENSE, VIT
Linear Transformation of Gaussian Random
Variables
The joint density function
109 Dr. [Link], Associate Professor, SENSE, VIT
12
21-01-2026
Tutorial Example
Zero-mean Gaussian random variables X1 and X2 having a covariance
matrix
110 Dr. [Link], Associate Professor, SENSE, VIT
Tutorial Example
SOLUTION
111 Dr. [Link], Associate Professor, SENSE, VIT
13
21-01-2026
Tutorial Example
1.074
112 Dr. [Link], Associate Professor, SENSE, VIT
Tutorial Example
Zero-mean Gaussian random variables X1 and X2 for which
113 Dr. [Link], Associate Professor, SENSE, VIT
14
21-01-2026
Tutorial Example
SOLUTION
114 Dr. [Link], Associate Professor, SENSE, VIT
Tutorial Example
115 Dr. [Link], Associate Professor, SENSE, VIT
15