Transform Methods: Chapter 3: Random Variables

You might also like

Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 17

Transform Methods

Chapter 3: Random Variables


EE-863 Stochastic Systems

College of Electrical & Mechanical Engineering


National University of Sciences & Technology (NUST)

Transform Methods

Logarithms are useful computational aids for performing


multiplications, using additions only

Transform methods held reduce computational effort e.g.


convolution may be performed in transform domain

F f1 ( x) f 2 ( x ) F1 ( )F2 ( )

Random Variables > Transform Methods

Characteristic Function

Random Variables > Transform Methods

Characteristic Function: Definition


X ( ) E e j X

Expected value of a
function of X

f X ( x)e j x dx

Fourier Transform of
the pdf

Inverse Fourier Transform of


Characteristic Function gives pdf
pdf and its characteristic function form a unique Fourier
transform pair
Random Variables > Transform Methods

Example 3.47 (Exponential Random


Variable)
X ( ) E e j X

f X ( x)e j x dx

pdf of exponential
random variable

Random Variables > Transform Methods

Characteristic Function for Discrete


Random Variable
X ( ) E e j X

f X ( x )e

j x

dx

Discrete random variable X

Integer-valued random variable X

Fourier Transform of the sequence


Random Variables > Transform Methods

Periodicity of Characteristic Function


Periodicity of discrete variable

j ( 2 ) k

j k

j 2 k

j k

.1 e

j k

Integer-valued random variable X

Fourier Transform of the sequence


periodic with period 2pi
Random Variables > Transform Methods

Recovering Probabilities from


Characteristic Function
Fourier Transform of the sequence
periodic with period 2pi
Inverse Fourier transform

Probabilities
Coefficients of Fourier Series of
periodic function
Random Variables > Transform Methods

Moment Theorem
X ( ) E e j X

f X ( x)e j x dx

Moments can be computed


from Characteristic Function

Proof:

Expand e j x into power series

If power series converges, pdf and characteristic function are completely


determined by moments of X
Random Variables > Transform Methods

Example 3.49 (Computing Mean and


Variance using Moment Theorem)
X ( ) E e j X

f X ( x)e j x dx

Differentiating once

Characteristic function
of exponential random
variable

Mean
Differentiating once again
Mean-square

Variance
Random Variables > Transform Methods

10

Probability Generating Function

Random Variables > Transform Methods

11

Probability Generating Function:


Definition
Expected value of a
function of N

GN ( z ) E z N

pN (k ) z k
k 0

N ( ) GN (e j )

z-Transform of the pmf

Relation between Characteristic


Function and Probability
Generating Function
j

ze
Random Variables > Transform Methods

12

Moment Theorem (PGF)


X ( ) E e j X

f X ( x)e j x dx

Moments can be computed


from Probability Generating
Function (PGF)

First derivative of PGF

Second derivative of PGF

Random Variables > Transform Methods

13

Moment Theorem (PGF) - continued


X ( ) E e j X

f X ( x) e j x dx

Mean

VAR[ N ] GN'' (1) GN' (1) (GN' (1)) 2

Random Variables > Transform Methods

Variance

14

Example 3.50 (Computing Mean and


Variance using Moment Theorem)
X ( ) E e j X

f X ( x)e j x dx

Probability Generating
Function of Poisson
Random Variable

Differentiating once

pmf of poisson random variable


Mean

Differentiating once again


Variance

Random Variables > Transform Methods

15

Laplace Transform of pdf

Random Variables > Transform Methods

16

Laplace Transform of pdf


Laplace Transform of
the pdf

X ( ) E e j X

f X ( x)e j x dx

Expected value of a
function of X

Moment Theorem

Random Variables > Transform Methods

17

You might also like