Week 9: Exercises

Date: November 10, 2024

Exercise 1. Let the continuous random variable X, Y have joint distribution


1/x if 0 < y < x < 1,

fX,Y (x, y) =
0

otherwise.
(a). compute E(X) and E(Y );
(b). compute the conditional pdf of Y given X = x, for all 0 < x < 1;
(c). compute Cov(X, Y ).

Solution 1.
(a).
Z ∞ Z ∞ Z 1Z x Z 1
1 1
E(X) = x · fX,Y (x, y)dydx = x · dydx = xdx =
−∞ −∞ 0 0 x 0 2
Z∞ Z ∞ Z 1Z x Z 1
1 1 1
E(Y ) = y · fX,Y (x, y)dydx = y · dydx = xdx =
−∞ −∞ 0 0 x 0 2 4
R∞ Rx 1
(b). The marginal pdf for X is fX (x) = −∞ fX,Y (x, y)dy = 0 x = 1 for 0 < x < 1 (and equals zero
otherwise). That is, X is uniform on the interval from 0 to 1. So the conditional pdf for Y given X is
1
fX,Y (y|x) = fX,Y (x, y)/fX (x) = , for 0 < y < x,
x
and 0 otherwise.
(c).
Z 1Z x Z 1
1 1 2 1
E(XY ) = xy · dydx = x dx = .
0 0 x 0 2 6
So,
1 1 1 1
Cov(X, Y ) = E(XY ) − E(X)E(Y ) = − · = .
6 2 4 24

Exercise 2. For any two random variables X and Y , show that

Var(X) = E(Var(X|Y)) + Var(E(X|Y )).

1
Solution 2. Note the fact that
E(X) = E(E(X|Y )). (1)
By definition, we have
 2 
Var(X) = E [X − E(X)]2 = E X − E(X|Y ) + E(X|Y ) − E(X)


 2   2    

= E X − E(X|Y ) + E E(X|Y ) − E(X) + 2E X − E(X|Y ) E(X|Y ) − E(X) .
The last term is equal to 0, which can be derived by (1)
  
   

E X − E(X|Y ) E(X|Y ) − E(X) = E E X − E(X|Y ) E(X|Y ) − E(X) |Y .
In the conditional distribution X|Y , X is the random variable. So in the expression,
  

E X − E(X|Y ) E(X|Y ) − E(X) |Y ,
E(X|Y ) and E(X) are constants. Thus,
  



E X − E(X|Y ) E(X|Y ) − E(X) |Y = E(X|Y ) − E(X) E(X − E(X|Y )|Y )



= E(X|Y ) − E(X) E(X|Y ) − E(X|Y )
= 0.
  

Thus, we have that E X − E(X|Y ) E(X|Y ) − E(X) |Y = E(0) = 0. Moreover, we see that
 2    2  
E X − E(X|Y ) = E E X − E(X|Y ) |Y = E(Var(X|Y ))
and
 2 
E E(X|Y ) − E(X) = Var(E(X|Y )),
which establishes the result.

Exercise 3. Let X be a random variable. Define the moment generating function of X, denoted by
MX (t) = E(etX ), provided that the expectation exists for t in some neighborhood of 0. If the moment
generating function exist and MX (t) = MY (t) for all t in the sam neighborhood of 0, then FX (u) = FY (u)
for all u.
(a). Compute MX (t) if X ∼ N (µ1 , σ12 );
(b). Compute MX+Y (t) if X ∼ N (µ1 , σ12 ), Y ∼ N (µ2 , σ22 ) and X, Y are independent;
(c). Show that X + Y ∼ N (µ1 + µ2 , σ12 + σ22 ) if X ∼ N (µ1 , σ12 ), Y ∼ N (µ2 , σ22 ) and X, Y are
independent.

Solution 3.

2
(a).
∞
(x − µ1 )2
Z
tX 1
MX (t) = E(e )=
exp(tx) p exp(− )dx
−∞ 2πσ12 2σ12
Z ∞
1 x2 − 2(µ1 + σ12 t)x + (µ1 + σ12 t)2 − σ14 t2 − 2µ1 σ12 t
= exp(− )dx
2σ12
p
−∞ 2πσ12
Z ∞
σ 2 t2 1 (x − (µ1 + σ12 t))2
= exp(µ1 t + 1 ) exp(− )dx
2σ12
p
2 −∞ 2πσ12
σ 2 t2
= exp(µ1 t + 1 ).
2
R∞ 1 (x−(µ1 +σ12 t))2
(The integral −∞ √ exp(− )dx = 1 since it is integral of the density function of
2 2πσ1 2σ 2 1

N (µ1 + σ12 t, σ12 ). )

(b).
MX+Y (t) = E(et(X+Y ) ) = E(etX · etY ) = E(etX )E(etY )
σ12 t2 σ 2 t2
) · exp(µ2 t + 2 )
= exp(µ1 t +
2 2
σ12 t2 σ22 t2
= exp(µ1 t + + µ2 t + ).
2 2
(c). To show X + Y ∼ N (µ1 + µ2 , σ12 + σ22 ), we just need to show MZ (t) = MX+Y (t) if Z ∼
N (µ1 + µ2 , σ12 + σ22 ).
Z ∞
tZ 1 (z − µ1 − µ2 )2
MZ (t) = E(e ) = exp(tz) p exp(− )dz
−∞ 2π(σ12 + σ22 ) 2(σ12 + σ22 )
Z ∞
1 x2 − 2(µ1 + µ2 + σ12 t + σ22 t)x + (µ1 + µ2 + σ12 t + σ22 t)2
= exp(− )
2(σ12 + σ22 )
p
−∞ 2π(σ12 + σ22 )
σ12 t2 σ 2 t2
· exp(µ1 t + + µ2 t + 2 )dx
2 2
∞
σ 2 t2 σ 2 t2 (x − (µ1 + µ2 + σ12 t + σ22 t))2
Z
1
= exp(µ1 t + 1 + µ2 t + 2 ) · exp(− )dx
2(σ12 + σ22 )
p
2 2 −∞ 2π(σ12 + σ22 )
σ12 t2 σ22 t2
= exp(µ1 t + + µ2 t + )
2 2
= MX+Y (t),
which proves the result.