Example
A device containing two key components fails when and only when both
components fail. The lifetime, T1 and T2, of these components are
independent with a common density function given by
e t
fT t
t0
otherwise
The cost, X, of operating the device until failure is 2T1 + T2. Find the
density function of X.
week 9
Convolution
Suppose X, Y jointly distributed random variables. We want to find the
probability / density function of Z=X+Y.
Discrete case
X, Y have joint probability function pX,Y(x,y). Z = z whenever X = x and
Y = z x. So the probability that Z = z is the sum over all x of these joint
probabilities. That is
p Z z p X ,Y x, z x .
x
If X, Y independent then
pZ z p X x pY z x .
x
This is known as the convolution of pX(x) and pY(y).
week 9
Example
Suppose X~ Poisson(1) independent of Y~ Poisson(2). Find the
distribution of X+Y.
week 9
Convolution - Continuous case
Suppose X, Y random variables with joint density function fX,Y(x,y). We want to find
the density function of Z=X+Y.
Can find distribution function of Z and differentiate. How?
The Cdf of Z can be found as follows:
FZ z P X Y z
z x
f x, y dydx
X ,Y
x y
f x, v x dvdx
X ,Y
x v
z
f x, v x dxdv.
X ,Y
v x
If
is continuous at z then the density function of Z is given by
f x, v x dx
XY
If X, Y independent then
fZ z
f x, z x dx
XY
f Z z f X x fY z x dx
x
This is known as the convolution of
fX(x) and fY(y).
Example
X, Y independent each having Exponential distribution with mean 1/. Find
the density for W=X+Y.
week 9
Some Recalls on Normal Distribution
If Z ~ N(0,1) the density of Z is
1
e
2
Z z
, z
If X = Z + then X ~ N(, 2) and the density of X is
f X x
z2
2
If X ~ N(, 2) then
Z
x 2
2 2
, x
X
~ N 0,1.
week 9
More on Normal Distribution
If X, Y independent standard normal random variables, find the density of
W=X+Y.
week 9
In general,
If X1, X2,, Xn i.i.d N(0,1) then X1+ X2++ Xn ~ N(0,n).
2
2
2
If X 1 ~ N 1 , 1 , X 2 ~ N 2 , 2 ,, X n ~ N n , n then
X 1 X 2 X n ~ N 1 n , 12 n2 .
If X1, X2,, Xn i.i.d N(, 2) then
Sn
2
Sn = X1+ X2++ Xn ~ N(n, n2) and X n n ~ N , n .
week 9
Sum of Independent 2(1) random variables
Recall: The Chi-Square density with 1 degree of freedom is the
Gamma( , ) density.
If X1, X2 i.i.d with distribution 2(1). Find the density of Y = X1+ X2.
In general, if X1, X2,, Xn ~ 2(1) independent then
X1+ X2++ Xn ~ 2(n) = Gamma(n/2, ).
Recall: The Chi-Square density with parameter n is
f X x
1
2
n
n/2
1
x
2
n
1
2
0 x
otherwise
Cauchy Distribution
The standard Cauchy distribution can be expressed as the ration of two
Standard Normal random variables.
Suppose X, Y are independent Standard Normal random variables.
Let Z Y . Want to find the density of Z.
X
week 9
10
Change-of-Variables for Double Integrals
Consider the transformation , u = f(x,y), v = g(x,y) and suppose we are
interested in evaluating F x, y dAxy .
Dxy
Why change variables?
In calculus: - to simplify the integrand.
- to simplify the region of integration.
In probability, want the density of a new random variable which is a
function of other random variables.
Example: Suppose we are interested in finding P A f X ,Y x, y dxdy
. .
A
Further, suppose T is a transformation with T(x,y) = (f(x,y),g(x,y)) = (u,v).
Then, P A f U ,V u, v dudv.
T A
Question: how to get fU,V(u,v) from fX,Y(x,y) ?
In order to derive the change-of-variable formula for double integral, we
need the formula which describe how areas are related under the
transformation T: R2 R2 defined by u = f(x,y), v = g(x,y).
week 9
11
Jacobian
Definition: The Jacobian Matrix of the transformation T is given by
f
x
J T x, y
g
x
y u , v
g x, y
y
The Jacobian of a transformation T is the determinant of the Jacobian
matrix.
In words: the Jacobian of a transformation T describes the extent to which
T increases or decreases area.
week 9
12
Change-of-Variable Theorem in 2-dimentions
Let x = f(u,v) and y = g(u,v) be a 1-1 mapping of the region Auv onto Axy
with f, g having continuous partials derivatives and det(J(u,v)) 0 on Auv.
If F(x,y) is continuous on Axy then
F x, y dxdy F f u, v , g u, v J u, v dudv
Axy
where
x
J u , v u
y
u
Auv
x
v 1
y J x, y
v
week 9
13
Example
Evaluate xydxdy where Axy is bounded by y = x, y = ex, xy = 2 and xy = 3.
Axy
week 9
14
Change-of-Variable for Joint Distributions
Theorem
Let X and Y be jointly continuous random variables with joint density
function fX,Y(x,y) and let DXY = {(x,y): fX,Y(x,y) >0}. If the mapping T given
by T(x,y) = (u(x,y),v(x,y)) maps DXY onto DUV. Then U, V are jointly
continuous random variable with joint density function given by
f X ,Y x u , v , y u , v J u , v
fU ,V u , v
if u , v DU ,V
otherwise
where J(u,v) is the Jacobian of T-1 given by
x
J u, v u
y
u
x
v
y
v
assuming derivatives exists and are continuous at all points in DUV .
week 9
15
Example
Let X, Y have joint density function given by
e x y
f X ,Y x, y
0
Find the density function of U
if x, y 0
otherwise
X
.
X Y
week 9
16
Example
Show that the integral over the Standard Normal distribution is 1.
week 9
17
Density of Quotient
Suppose X, Y are independent continuous random variables and we are
interested in the density of Z Y .
X
y
z
, wx .
Can define the following transformation
x
The inverse transformation is x = w, y = wz. The Jacobian of the inverse
transformation is given by
x
J w, z w
y
w
x
z 1 0 w
y z w
z
Apply 2-D change-of-variable theorem for densities to get
fW ,Z w, z f X ,Y w, wz w f X w fY wz w
The density for Z is thengiven by
f Z z f X w fY wz w dw
week 9
18
Example
Y
Suppose X, Y are independent N(0,1). The density of Z
is
X
week 9
19
Example F distribution
X /n
Z
.
Y /m
Suppose X ~ (n) independent of Y ~
This is the Density for a random variable with an F-distribution with
parameters n and m (often called degrees of freedom). Z ~ F(n,m).
(m)
week 9
. Find the density of
20
Example t distribution
Suppose Z ~ N(0,1) independent of X ~ (n). Find the density of T
2
Z
X
.
n
This is the Density for a random variable with a t-distribution with
parameter n (often called degrees of freedom). T ~ t(n)
week 9
21
Some Recalls on Beta Distribution
If X has Beta(,) distribution where > 0 and > 0 are positive
parameters the density function of X is
1
1
x 1 x
0 x 1
f X x
otherwise
If = = 1, then X ~ Uniform(0,1).
If = = , then the density of X is
f X x x 1 x
for 0 x 1
otherwise
Depending on the values of and , density can look like:
If X ~ Beta(,) then E X
.
and V X
2
1
week 9
22
Derivation of Beta Distribution
Let X1, X2 be independent 2(1) random variables. We want the density of
X1
X1 X 2
Can define the following transformation
X1
Y1
,
Y2 X 1 X 2
X1 X 2
week 9
23
