INTERMEDIATE MATHEMATICAL STATISTICS I

HOMEWORK 4

Assigned September 13, 2022, due September 20, 2022

This homework pertains to materials covered in Lecture 7 and 8. The
assignment can be typed or handwritten, with your name on the document,
and with properly labeled computer output for those problems that require
computer simulations. To obtain full credit, please write clearly and show
your reasoning. If you choose to collaborate, the write-up should be your
own. Please show your work! Upload the file to the Week 4 Assignment on
Canvas or hand it in before class.
Problem 1 (10 points). Problem 2.2 of Casella & Berger.
Solution.
(a) Since g(x) = x2 is monotone in x ∈ (0, 1), we get the inverse function
√
g −1 (y) = y for y ∈ (0, 1).
Thus the PDF of Y can be derived as follows:
 −1 
 dg (y)  1
−1
fY (y) = fX (g (y))  = √ 1(0, 1).
dy  2 y
(b) Since the support of X is (0, 1), the support of Y is (0, ∞). Also,
g −1 (y) = exp(−y) and the PDF of Y is
 −1 
−1
 dg (y) 
fY (y) = fX (g (y))   = fX (exp(−y)) exp(−y)
dy 
(n + m + 1)! −(n+1)y
= e (1 − e−y )m 1(0, ∞).
n!m!
(c) Since g −1 (y) = log y and y ∈ (1, ∞),
 −1 
 dg (y) 
−1
fY (y) = fX (g (y))   = fX (log y) 1
dy  y
1 log y − log2 y/(2σ2 )
= 2 e 1(1, ∞)
σ y
log y
= 2 y − log y/(2σ )−1 1(1, ∞).
2

σ
Problem 2 (10 points; Weibull distribution). Let X denote a random vari-
able with CDF
FX (x) = 1 − exp(−x), x > 0.
Let Y = X 1/θ where θ > 0. Find the CDF and PDF of Y .
1
 INTERMEDIATE MATHEMATICAL STATISTICS I HOMEWORK 4 2

Solution. By definition, the CDF of Y is

FY (y) = P (Y ≤ y) = P (X 1/θ ≤ y)
= P (X ≤ y θ ) = FX (y θ ) = 1 − exp(−y θ ), y > 0.
The PDF of Y is
dFY (y)
= θy θ−1 e−y 1(0, ∞).
θ
fY (y) =
dy
Problem 3 (10 points). Problem 2.3 of Casella & Berger.
Solution. The values where Y has positive probability mass are 0, 21 , 23 , 34 , 45 , . . ..
The inverse of the transformation is g −1 (y) = 1−y
1
− 1. Therefore the PMF
of Y is
1 2 1/(1−y)−1 1 2 1/(1−y)
   
1 2 3 4
fY (y) = = , y = 0, , , , , . . .
3 3 2 3 2 3 4 5
Problem 4 (10 points). Let X be a real-valued random variable (i.e., one-
dimensional random variable) with PDF p. Find the PDF of Y = |X|.
Solution. First let’s derive the CDF of Y :
(
0, if y < 0,
FY (y) = P (Y ≤ y) = P (|X| ≤ y) =
P (−y ≤ X ≤ y) = FX (y) − FX (−y), if y ≥ 0.

Thus the PDF is

dFY (y)
fY (y) = = {p(y) + p(−y)}1(0, ∞).
dy
Problem 5 (5 points). Problem 2.6 of Casella & Berger.
Solution.
(a) The CDF is
(
0, if y < 0,
FY (y) = P (Y ≤ y) = P (|X|3 ≤ y) =
FX (y 1/3 ) − FX (−y 1/3 ), if y ≥ 0.

Thus the PDF is

dFY (y)
fY (y) = = {fX (y 1/3 )y −2/3 /3 + fX (−y 1/3 )y −2/3 /3}1(0, ∞)
dy
1
= y −2/3 e−y 1(0, ∞).
1/3

3
To verify that the PDF integrates to 1,
Z ∞ Z ∞ ∞
1 −2/3 −y1/3 −y 1/3 

fY (y)dy = y e dy = −e  = 1.
−∞ 0 3 0
 INTERMEDIATE MATHEMATICAL STATISTICS I HOMEWORK 4 3

(b) The CDF is

(
0, if y ̸∈ (0, 1),
FY (y) = P (Y ≤ y) = P (1−X 2 ≤ y) = √ √
P (X ≥ 1 − y) + P (X ≤ − 1 − y), if y ∈ (0, 1).
Note
p p p p
P (X ≥ 1 − y) + P (X ≤ − 1 − y) = 1 − FX ( 1 − y) + FX (− 1 − y).
Therefore the PDF can be derived:
dFY (y) p 1 p 1
fY (y) = = fX ( 1 − y) √ + fX (− 1 − y) √
dy 2 1−y 2 1−y
√ 2
√ 2
3( 1 − y + 1) + 3(− 1 − y + 1)
= √ 1(0, 1).
16 1 − y
To verify that the PDF integrates to 1,
Z ∞ Z 1 √ √
3( 1 − y + 1)2 + 3(− 1 − y + 1)2
fY (y)dy = √ dy
−∞ 0 16 1 − y
√ Z 1
s= 1−y 3(s + 1)2 + 3(−s + 1)2
= ds
0 8
Z 1 1
3(s2 + 1) s3 + 3s 
= ds =  = 1.
0 4 4 
0
(c) By the law of total probability,
P (Y ≤ y) = P (Y ≤ y|X ≤ 0)P (X ≤ 0) + P (Y ≤ y|X > 0)P (X > 0)
= P (1 − X 2 ≤ y|X ≤ 0)P (X ≤ 0) + P (1 − X ≤ y|X > 0)P (X > 0)
p
= P (X ≤ − 1 − y|X ≤ 0)P (X ≤ 0) + P (X ≥ 1 − y|X > 0)P (X > 0)
p p
= P (X ≤ − 1 − y) + P (X ≥ 1 − y) = FX (− 1 − y) + 1 − FX (1 − y).
Therefore the PDF can be derived:
dFY (y) p 1
fY (y) = = fX (− 1 − y) √ + fX (1 − y)
dy 2 1−y
√
3(− 1 − y + 1)2 3
= √ + (2 − y)2 , 0 < y < 1.
16 1 − y 8
To verify that the PDF integrates to 1,
Z ∞ Z 1 √
−3( 1 − y + 1)2 3
fY (y)dy = √ + (2 − y)2 dy
−∞ 0 16 1 − y 8
Z 1 1
3(−s + 1)2 (2 − y)3 
= ds −
8 8

0 
0
1
(−s + 1)3  1
=−  − + 1 = 1.
8  8
0
 INTERMEDIATE MATHEMATICAL STATISTICS I HOMEWORK 4 4

Problem 6 (15 points). Problem 2.7 of Casella & Berger.

Solution.
(1) The CDF is
( √
2 P (1 ≤ X ≤ y), if y ∈ (1, 4),
FY (y) = P (Y ≤ y) = P (X ≤ y) = √ √
P (− y ≤ X ≤ y), if y ∈ (0, 1).
( √
F ( y) − F (1), if y ∈ (1, 4),
= √ √
F ( y) − F (− y), if y ∈ (0, 1).
Therefore the PDF is
( √ 1
f ( y) 2√ y, if y ∈ (1, 4),
fY (y) = √ 1 √ 1
f ( y) 2√y + f (− y) 2√ y, if y ∈ (0, 1).
( √y+1
√ , if y ∈ (1, 4),
= 92 y
9 y,
√ if y ∈ (0, 1).
(2) Theorem 2.1.8 is not directly applicable because the image or the
set Y = {y : y = gi (x) for x ∈ Ai } is NOT the same for A1 = (−1, 0)
and A2 = (0, 2).
However, if we partition the range of Y into B1 , B2 , . . . such that
we can partition the pre-image of each Bi into monotone parts which
all have a image of Bi . That is, we can apply Theorem 2.1.8 within
each Bi , which leads to the result in (a).
Problem 7 (10 points). Problem 2.9 of Casella & Berger.
Solution. We want Y to be U (0, 1), which means for y ∈ (0, 1)
y = P (Y ≤ y) = P (X ≤ u−1 (y)) = FX (u−1 (y)).
1(1, 3), the inverse function can be calculated: FX−1 (y) =
2
Since FX (x) = (x−1)
√ 4
4y + 1. This leads to
u−1 (y) = FX−1 (y) = ( 4y + 1)1(0, 1)
p

Invert u−1 again to get u(x), and we have

u(x) = FX (x).
Problem 8 (10 points). Suppose X = (X1 , X2 ) has PDF
p(x1 , x2 ) = x−2 −2
1 x2 , x1 > 1, x2 > 1.
Find the PDF of X1 X2 .
Solution. We first define a transformation
(
U = X1 X2 ,
V = X2 ,
 INTERMEDIATE MATHEMATICAL STATISTICS I HOMEWORK 4 5

whose inverse is (
X1 = U/V,
X2 = V.
Then the support of the transformed random vector is {(u, v) : 1 < v < u},
and the Jacobian determinant is
 
1/v −u/v 2  1
J =  = .
0 1  v
Thus the joint PDF of (U, V ) is
1
f (u, v) = p(u/v, v)|J|1(1 < v < u) = 1(1 < v < u).
u2 v
Lastly, we can derive the marginal PDF of U = X1 X2 via
Z ∞ Z u
1 log u
fU (u) = f (u, v)dv = 2
dv = 2 1(u > 1).
0 1 u v u
Problem 9 (12 points). Let X1 and X2 denote independent, real-valued
random variables with absolutely continuous distributions with PDFs p1
and p2 , respectively. Let Y = X1 /X2 . Prove that Y has PDF
Z ∞
pY (y) = |z|p1 (zy)p2 (z)dz.
−∞

Solution. We first define a transformation

(
Y = X1 /X2 ,
Z = X2 ,
whose inverse is (
X1 = Y Z,
X2 = Z.
Thus the Jacobian determinant is
 
z y 
J =
  = z,
0 1
and the joint PDF of (Y, Z) is
f (y, z) = p1 (yz)p2 (z)|J| = p1 (yz)p2 (z)|z|.
The marginal PDF of Y can be derived:
Z ∞ Z ∞
pY (y) = f (y, z)dz = |z|p1 (zy)p2 (z)dz.
−∞ −∞

Problem 10 (8 points). Let X1 , X2 denote independent random variables

such that Xj has an absolutely continuous distribution with density function
λj exp(−λj x), x > 0,
j = 1, 2, where λ1 > 0 and λ2 > 0. Find the density of Y = X1 /X2 .
 INTERMEDIATE MATHEMATICAL STATISTICS I HOMEWORK 4 6

Solution. From the previous the problem, we know the marginal PDF
for Y is
Z ∞
pY (y) = |z|λ1 exp(−λ1 zy)λ2 exp(−λ2 z)dz
0
Z ∞
= λ1 λ2 z exp{−(λ1 y + λ2 )z}dz
0
∞
λ1 λ2 
= 2
[−(λ1 y + λ2 )z exp{−(λ1 y + λ2 )z} − exp{−(λ1 y + λ2 )z}]
(λ1 y + λ2 ) 0
λ1 λ2
= , y > 0.
(λ1 y + λ2 )2
Problem 11 (Bonus 10 points). Let X denote a random variable with a
PDF:
 d1 /2  −(d1 +d2 )/2
1 d1 d1
fX (x) =   x d1 /2−1
1+ x 1(x > 0),
B d21 , d22 d2 d2

in which 1(·) is the indicator function, B(·, ·) is the Beta function, and
d1 , d2 > 0 are fixed positive constants.
Find the PDF of
d1 X
Y = .
d2 1 + (d1 /d2 )X
(d1 /d2 )x
Solution. The transformation g(x) = 1+(d1 /d2 )x is an increasing function
(d1 /d2 )
because g ′ (x) = {1+(d1 /d2 )x}2
> 0 for all x > 0. Also, g maps (0, ∞) to (0, 1)
(d1 /d2 )∞
as g(∞) = 1+(d 1 /d2 )∞
= 1.
y
Since g −1 (y) = dd21 1−y , the
density function of Y is
 −1 
 dg (y) 
fY (y) = fX (g −1 (y))  
dy 
 d1 /2  d1 /2−1  −(d1 +d2 )/2
1 d1 d2 y y d2 1
=   1+
B d21 , d22 d2 d1 1 − y 1−y d1 (1 − y)2

1
=   y d1 /2−1 (1 − y)d2 /2−1 , 0 < y < 1.
d1 d2
B 2 , 2

That is, Y is a Beta distribution with parameters d1 /2 and d2 /2.