ELEC2600 Tutorial 7
Functions/Transforms of a Random Variable
Functions/Transforms of a Random Variable
2
Problem Statement
Given a random variable X with known distribution and a real
valued function g(x), such that Y = g(X) is also a random variable.
Find the distribution of Y.
Solution methodology
Idea: find an equivalent events in X for suitably defined events in Y.
If X is discrete, Y must be discrete, irrespective of g(x).
If X is continuous, Y may be discrete, continuous or mixed-type.
This depends on the form of g(x).
Discrete/Continuous X, Discrete Y = g(X)
3
The pmf at the possible values of Y:
Example 1
Discrete/Continuous X, Discrete Y = g(X) Example 1
Solution
otherwise
otherwise
4
Discrete/Continuous X, Discrete Y = g(X) Example 1
Solution
otherwise
5
Discrete/Continuous X, Discrete Y = g(X) Example 2
X is a continuous random variable which takes value between interval
[-1,1], Y=g(X) is defined as follows:
a) The cdf of X is given below. Find the pmf of Y.
b) If X follows uniform distribution, find the pmf of Y.
6
Discrete/Continuous X, Discrete Y = g(X) Example 2
Solution
This problem shows how to do the transform from continuous random
variable to discrete random variable.
Approach:
(1) Find out the range of Y:
(2) Y can be presented as:
The range of Y is: {-1,0,1}
(3) Find equivalent events
in X for suitably defined
events in Y
7
Discrete/Continuous X, Discrete Y = g(X) Example 2
Solution
a)
0.5, y = −1
0.45, y=0
pY ( y ) =
0.05, y =1
0, otherwise
b)
8
Discrete/Continuous X, Discrete Y = g(X) Example 3
Solution
0.25, y = −1
0.5, y=0
pY ( y ) =
0.25, y =1
0, otherwise
9
Continuous X, Continuous Y = g(X)
10
CDF of Y
PDF of Y
Leibniz Integral Rule
Continuous X, Continuous Y = g(X) Example 1
Solution
y +1
FY ( y )= P [Y ≤ y ]= P X − 1 ≤ y = P − y + 1 ≤ X ≤
2
y + 1 = ∫ f X ( x ) dx
− y +1
fY ( y= (
) fx )( y + 1) − f ( − y + 1)( − )
' '
y +1 x y +1
f ( x y + 1) f ( − y + 1) γ e
x
−γ y +1
= + =
2 y +1 2 y +1 2 y +1
11
Continuous X, Continuous Y = g(X) Example 2
The amplitude of a radio signal 𝑋𝑋 is a Rayleigh random variable with
PDF:
(a) Find the PDF of 𝑍𝑍=(𝑋𝑋−1)+. Here, (𝑋𝑋−1)+=max {𝑋𝑋−1,0}.
(b) Find the PDF of 𝑍𝑍=𝑋𝑋2.
12
Continuous X, Continuous Y = g(X) Example 2
13