Homework # 7, #8

14.7. Assume that X (u) is real. Then

X (u) = X (u) = X (u) = X (u)

d
Therefore X = X, i.e., X is symmetric.
d
Assume X = X. Then

X (u) = X (u) = X (u)

So X (u) is real.

14.8. Notice that

XY (u) = X (u)Y (u) = X (u)X (u) = |X (u)|2

is a real function. So X Y is symmetric.

15.5. Write

X
SN = Sn 1{N=n}
n=0

We have

X  m
X  m
X 
ESN = E Sn 1{N=n} = E lim Sn 1{N=n} = lim E Sn 1{N=n}
m m
n=0 n=0 n=0
m m
(1)
X X X
= lim ESn 1{N=n} = lim E(Sn )P {N = n} = E(Sn )P {N = n}
m m
n=0 n=0 n=0

where the third equality follows from dominated convergence with the controaling random
veriable

X
Y = |Sn |1{N=n}
n=1

This is because of
 m
X 


 Sn 1 {N=n}
Y
 m = 1, 2,
n=0

and

X nX
X n o


EY = E |Sn |1{N=n} E|Xj |1{N=n}
n=1 n=1 j=1

(2)
X

X

= n E|Xj | P {N = n} = E|X1 | nP {N = n} = E|X1 | EN <
n=1 n=1

1
Inaddition, the fifth equality in (1), and the third step in (2) follows from independence
between {Xj } and N .

Further, from (1) we have

X


ESN = n EX1 P {N = n} = EX1 EN
n=0

This is a special form of Walds first identity.

15.13. Write Z = (X, Y1 , , Yn )

n n
X o
Z (u0 , u1 , , un ) = E exp iu0 X + i uk Y k
k=1

Notice that
n
X  n
X  n
X
u0 X + uk Y k = u0 uk X + uk X k
k=1 k=1 k=1
n n n
1 X X X
= u0 uk Xk + uk X k
n
k=1 k=1 k=1
n  n 
X 1 X 
= u0 uk + uk X k
n
k=1 k=1

Z (u0 , u1 , , un )
n n 2 n
2 1 
    
Y X  1 X 
= exp u0 uk + uk + i u0 uk + uk
2 n n
k=1 k=1 k=1
 2 X n   n 2
1 X 
= exp u0 uk + uk
2 n
k=1 k=1
n   n 
X 1 X 
+ i u0 uk + uk
n
k=1 k=1

n   n 
X 1 X 
u0 uk + uk = u0
n
k=1 k=1

2
 n   n 2
X 1 X 
u0 uk + uk
n
k=1 k=1
n
X 1   Xn 2 2  n
X  
2
= u0 uk + uk u0 uk + uk
n2 n
k=1 k=1 k=1
n n n n
1 X 2 2  X  X  X
= u0 uk + uk u0 uk + u2k
n n
k=1 k=1 k=1 k=1
 n n 2 n n
1 X  X 1 X
  2 X
= u0 uk + uk uk + u2k
n n
k=1 k=1 k=1 k=1
n n
1 2 X 1X 2
= u + u2k uk
n 0 n
k=1 k=1

Therefore,

Z (u0 , u1 , , un )
n 2  n n 2 
2 X 2 1  X 2
o 
2
= exp u iu0 exp uk uk
2n 0 2 n
k=1 k=1

Letting u0 = 0, we obtain the characteristic function of Y = (Y1 , , Yn )

 n n 2 
2 X 2 1  X 2

Y (u1 , , un ) = Z (0, u1 , , un ) = exp uk uk
2 n
k=1 k=1

Similarly, letting u1 = = un = 0 we obtain the characteristic function of X

n 2 2 o
X (u0 ) = Z (u0 , 0, , 0) = exp u0 iu0
2n

In particular X N (, 2 /n).
In addition,
Z (u0 , u1 , , un ) = X (u0 )Y (u1 , , un )
This implies that X and (Y1 , , Yn ) are independent. Since
n
1X 2
Sn2 = Y
n j=1 j

is a function of (Y1 , , Yn ), X and Sn2 are independent.

Alternative solusion. We may also use Gaussian property to give a proof. First, we
claim that (x, Y1 , , Yn ) is Gaussian. Indeed, any linear combination of x, Y1 , , Yn is
1-dimensional normal, as it can be written as a linear combination of X1 , , Xn .

3
 Second,
n
 
2 1 1 X
Cov(x, Yj ) = E (x )(Xj ) E x = Var(Xj ) 2 Var(Xk ) = 0
n n
k=1

for j = 1, , n. By Theorem 16.4, x and (Y1 , , Yn ) are independent. Consequently, x

and S 2 are independent.

16.2. Let A (, ) be Borel-measurable.

o o
P {Z A} = P {Y A, |Y | a + P {Y A, |Y | > a

By the symmetry of Y , we can replace Y by Y in the second ter on the right hand side:
o o o
P {Y A, |Y | > a = P {(Y ) A, | Y | > a = P {Y A, |Y | > a

Thus,
o o
P {Z A} = P {Y A, |Y | a + P {Y A, |Y | > a = P {Y A}

d
Hence, Z = Y .
Remark. What we really need here is not the normality, but symmetry of Y .

16.7. By definition of Gaussian random variable, Y N (, 2 ) where

n
X
= EY = aj E(Xj )
j=1

and
n
2 hX
i2
2 = Var(Y ) = E Y EY = E

aj Xj E(Xj )
j=1
n
X 

2 X
a2j



=E Xj E(Xj ) +2 aj ak Xj E(Xj ) Xk E(Xk )
j=1 j<k
n
X X
= a2j Var(Xj ) + 2 aj ak Cov(Xj , Xk )
j=1 j<k

16.16. We need only to exam two things: First, (X, Y X) is 2-dimensional Gaussian.
Second, Cov(X, Y X) = 0.

4
 For any constant a1 , a2 ,

a1 X + a2 (Y X) = (a1 )X + a2 Y

Since (X, Y ) is Gaussian, a1 X + a2 (Y X) is normal. This shows that (X, Y X) is

Gaussian.
By linearty,
2
Cov(X, Y X) = Cov(X, Y ) Cov(X, X) = hboxCov(X, Y ) X
2
By the fact X = Y2 q q
2 2
X = X Y2 = Cov(X, Y )
Hence,
Cov(X, Y X) = Cov(X, Y ) Cov(X, Y ) = 0.

17.3. By Chebyshev inequality, for any > 0

n
n 1 X n n
 o 1 1 X  1 X 1
P  Xj  2 V ar Xj = 2 2 V ar(Xj ) = 2
 
n n n n
j=1 j=1 j=1

The right hand side tens to 0 as n . So we have

n
n 1 X  o
lim P  Xj  = 0
 
n n
j=1

17.4 Replace n by n2 in above estimate.

n 1 Xn2  o 1
P  2 Xj  2 2
 
n j=1 n

Hence,
n 1 Xn2
X  o X 1
P  2 Xj  <
 
n=1
n j=1 n=1
n 2
2

By Borel-Cantelli lemma,
\
 [ n n2 o
1 X 

P  2 Xj  =0
m=1 n=m
n j=1

Notice that
 1 X n2  \ n n2
1 X 
 [  o
lim sup  2 Xj  = Xj 
 
 2
n n j=1 m=1 n=m
n j=1

5
Thus,
1 X n2 
lim sup  2 Xj  a.s.
 
n n j=1

Letting 0+ on the right hand side we have

2
n
1 X
lim Xj = 0 a.s.
n n2
j=1