Math 510 HW #19 selected answers

Problem 1 Let X be a random variable with the following mass distribution:

0 .3
1 .5
2 .2
Find the moment generating function for X.

MX (t) = E[etX ]
= .3 e0t + .5 e1t + .2 e2t
= .3 + .5 et + .2 e2t
Problem 2 Let X be a random variable with the density function

x, 0 < x < 1
f (x) =
0, otherwise
Find the moment generating function for X.

M (t) = E[etX ]
Z
= f (x) etx dx

Z 1
= x etx dx
0
Z 1
x tx x=1 1 tx
= e  e dx
t x=0 0 t
1 t 1
= e 2 (et 1)
t t
Problem 3 Suppose you have a random variable X whose moments are given
by E[X n ] = n!. Find the moment generating function for X.

X E[X n ] n
M (t) = t
n=0
n!

X n! n
= t
n=0
n!

X
= tn
n=0
1
=
1t
Problem 4 Prove that the sum of two independent Poisson random variables
with = 1 is a Poisson random variable with = 2 using moment generating
functions.

The moment generating function for a Poisson random variable with = 1 is

t t
MX (t) = ee 1 . The sum of two independent such is MX (t)2 = e2(e 1) , which
is the moment generating function for a Poisson random variable with = 2.
By the uniqueness of moment generating functions, the sum of two independent
Poisson random variables with = 1 is a Poisson random variable with = 2.

Problem 5 Find the first four moments (i.e., E[X] through E[X 4 ]) of an ex-
ponential distribution with parameter , by taking derivatives of the moment
generating function.

The moment generating function for an exponential random variable is

M (t) =
t
The first four derivatives are:

M 0 (t) =
( t)2
2
M 00 (t) =
( t)3
6
M 000 (t) =
( t)4
24
M 0000 (t) =
( t)5

and to find the moments, we plug in t = 0:

1
E[X] = M 0 (0) =

2
E[X 2 ] = M 00 (0) = 2

6
E[X 3 ] = M 000 (0) = 3

24
E[X 4 ] = M 0000 (0) = 4

Problem 6 Suppose the moment generating function for X is
2
M (t) = e4t +3t
.

Describe what kind of random variable X is, with what values of the parameters.
 By the uniqueness of moment generating functions, X must be normal, since
normal random variables have moment generating functions
2 2
e 2 t +t .

Matching terms, we conclude 2 /2 = 4 so that = 8, and = 3.

Problem 7 Theoretical ex. #46:

We expand the moment generating function for Z:
2
MZ (t) = et /2
 j
X 1 t2
=
j=0
j! 2

X 1 t2j
=
j=0
j! 2j

The tn term is 0 if n is odd, and when n is even, we write n = 2j and the term
is
1 n
t .
2j j!
Comparing this with
n n
t
n!
we get that when n = 2j,
(2j)!
n = .
2j j!
Theoretical ex. #49:
The fact that log X is normal means
2 t2
E[et log X ] = e 2 +t

and
E[et log X ] = E[X t ]
so setting t = 1 and t = 2 we get
2
E[X] = E[X 1 ] = e 2 +

and
4 2
E[X 2 ] = e 2 +2

so that 2 2
V ar(X) = E[X 2 ] E[X]2 = e2 +2
e +2

Theoretical ex. #50:

 (t) = log M (t)
M 0 (t)
0 (t) =
M (t)
M 00 (t) M (t) M 0 (t) M 0 (t)
00 (t) =
M (t)2
M (0) M (0) M 0 (0) M 0 (0)
00
00 (0) =
M (0)2
E[X ] 1 E[X]2
2
=
12
= E[X ] E[X]2
2

= V ar(X)