Institute of Technology of Cambodia

Miterm Exam (Type 2)

Class: Engineering Year III
Subject: Statistics Duration: 3h :00 mn
December 25th, 2019

Remark: Calculators are allowed.

1. Let X1 , X2 , ... , Xn be a random sample from a population X with pdf


 1 θ1 −1
x , if 0 < x ≤ 1
f (x; θ) = θ
0, otherwise.

where θ > 0 is an unknown parameter.

(1.1) Let Y = − ln X. Show that Y ∼ Exp(θ).
(1.2) Find the MLE θ̂n for θ. Is θ̂n an efficient estimator of θ?
2nθ̂n
(1.3) Let U = . Find the mgf of U and deduce that U ∼ χ2 (2n).
θ
ν 1
[Recall that V ∼ χ2 (ν) ⇔ Mν (t) = (1 − 2t)− 2 , t < ]
2
(1.4) Derive a 100(1 − α)% confidence interval for θ.
(1.5) Find the best critical region for testing H0 : θ = 1 versus Ha : θ = θa , where θa > 1
when α = 0.01 and n = 15.
(1.6) Is the test in part (e) a UMP test for testing H0 : θ = 1 vs Ha : θ > 1? Justify your
answer?

2. Let X1 , X2 , ... , X15 be a random sample of size 15 from N(µ, σ 2 ).

15  2
X Xi − µ
(2.1) What is the distribution of the statisitc V = ?
i=1
σ
(2.2) Find the 95th percentile of V.
3. An instructor has just given an exam on Statistics to his students. Below is a random
sample of scores of 15 students:

85, 90, 70, 50, 45, 60, 80, 75, 52, 35, 68, 95, 86, 88, 65

Assume that the score is normally distributed. [Hint: x̄ = 69.6, s = 18.24]

(3.1) Construct a 95% confidence interval for the true average µ of the score.
(3.2) The instructor claims that the true variance of the scores of this statistics test is
less than 300. Is there enough evidence to support the instructor’s claim ? Use
α = 0.05.

1
4. Let X1 , X2 , ... , Xn be a random sample from a distribution with mean µ, and unknown
n
X
variance σ 2 < ∞ and let a1 , a2 , ... , an be real numbers such that ai = 1. Define X̄ =
i=1
n
X
ai Xi .
i=1

(4.1) Show that X̄ is an unbiased estimator of µ.

n
X
(4.2) Show that V(X̄) ≤ V(X̄) (hence among all estimators of µ of the form ai Xi ,
i=1
n
X
where ai = 1 is the MVUE).
i=1

2
3
 Solution

1. (1.1) Show that Y = − ln X ∼ Exp(θ)

y = − ln x ⇔ x = e−y
dx
⇒J= = −e−y
dy
⇒ |J| = e−y
Thus, the pdf of Y is

 1 −y θ1 −1
−y −y (e ) , if y ≥ 0
g(y) = f (e ) · |J| = e θ
0, otherwise

 1 − θy
e , if y ≥ 0
= θ
0, otherwise

Thus, Y ∼ Exp(θ).
(1.2) Find the MLE of θ
n n n
! θ1 −1
Y Y 1 1
−1
Y
L(θ) = f (x; θ) = xi
θ
= θ−n xi
i=1 i=1
θ i=1
 Xn
1
ln L(θ) = −n ln θ + −1 ln xi
θ i=1
n
∂ n 1 X
ln L(θ) = − − 2 ln xi
∂θ θ θ i=1
n
1X
⇒θ=− ln xi = ȳ
n i=1
n
1X
Thus the MLE of θ is θ̂n = − ln Xi = Ȳ.
n i=1
Is θ̂n an efficient estimator?
n
! n
1X 1X nθ
E(θ̂n ) = E Yi = E(Yi ) = =θ (1)
n i=1 n i=1 n
n
! n
1X 1 X nθ2 θ2
V(θ̂n ) = V Yi = 2 V(Yi ) = 2 =
n i=1 n i=1 n n

Since X(Ω) = [0, ∞), then

 
∂
I(θ) = V ln f (x; θ)
∂θ
 
1
Since ln f (x; θ) = − ln θ + − 1 ln x then
θ
∂ 1 1
ln f (x; θ) = − − 2 y
∂θ θ θ
4
 then
1 θ2 1
I(θ) =4
V(Y) = 4
= 2
θ θ θ
2
1 θ
⇒ = = V(θ̂n ) (2)
nI(θ) n

From (1) and (2), we have θ̂n is an efficient estimator of θ.

n
2nθ̂n 2X
(1.3) Find the mgf of U = =− ln Xi
θ θ i=1
 2t Pn 
MU (t) = E etU = E e− θ i=1 ln Xi


h  2t Pn in h  2t in
−θ
= E e i=1 ln X
= E X− θ

Since
1
1 1 −1 1 1 1−2t
  Z Z
− 2t − 2t
E X θ = x θx θ dx = x θ dx
0 θ θ 0
" 1−2t #1
1 x θ 1
= 1−2t
= (1 − 2t)−1 , t <
θ θ
2
0

2n 1
So, MU (t) = (1 − 2t)−n = (1 − 2t)− 2 , t < .
2
Hence, U ∼ χ2 (2n).
(1.4) Find 100(1 − α)% CI for θ
Since U ∼ χ2 (2n), by symmetry principle, we have
 
1 − α = P χ21− α ,2n ≤ U ≤ χ2α ,2n
2 2
!
2nθ̂n
= P χ1− α2 ,2n ≤ ≤ χ α2 ,2n
θ
!
2nθ̂n 2nθ̂n
=P ≤θ≤
χ2α ,2n χ1− α2 ,2n
2

Thus a 100(1 − α)% CI for θ is

" #
2nθ̂n 2nθ̂n
I(θ) = , .
χ2α ,2n χ1− α2 ,2n
2

(1.5) H0 : θ = 1 vs Ha : θ = θa , θa > 1
By Neyman-Pearson lemma, we have, for k > 0,
L(θ0 ) L(1)
≤k⇔ ≤k
L(θa ) L(θa )
1
⇔
θ 1−1 ≤ k
n
θa−n
Q
x
i=1 i
a

n
! θ 1−1
Y a

⇔ xi ≤ kθa−n
i=1

5
   n
1 X
ln xi ≤ ln kθa−n


⇔ 1−
θa i=1
n
θa
ln kθa−n ∼
X

⇔ ln xi ≤ =c
i=1
θa − 1
( n
)
X
Thus, RR = (x1 , · · · , xn ) : ln xi ≤ c where the constant c is defined by
i=1

n
!
X
α=P ln Xi ≤ c|θ = 1
i=1
 
2c
= P U ≥ − |θ = 1 , U ∼ χ2 (2n)
θ
= P(U ≥ −2c)

So, −2c = χ2α,2n = χ20.01,30 = −50.892 ⇒ c = −25.446.

( 15
)
X
Thus, RR = (x1 , · · · , xn ) : ln xi ≤ −25.446 .
i=1
n
X
(1.6) Since the test H0 : θ = 1 vs Ha : θ = θa defines the test statistic ln Xi and RR do
i=1
not depend on θa , for each θa > 1, then it is a UMP test for testing H0 : θ = 1 vs
Ha : θ > 1.
2. (2.1) The distribution of V is χ2 (15).
(2.2) Let a be the 95-th percentile of V. We have

0.05 = P(V ≥ a)

Thus, a = χ20.05,15 = 24.996.

3. (3.1) A 95% CI for µ is
h s si
I(µ) = x̄ − t · , x̄ + t 2 ,n−1 ·
α α
,n−1
 n 2
 n
18.24
= 69.6 ± t0.025,14 · √
15
 
18.24
= 69.6 ± 2.145 · √
15
= [59.498, 79.702].

(3.2) Test H0 : σ 2 = 300 vs Ha : σ 2 > 300

Test statistic value
(n − 1)s2 14 × 18.242
χ2 = = = 15.53
σ02 300

RR = χ2 : χ2 ≥ χ2α,n−1 = χ0.05,14 = 23.685



Since χ2 = 15.53 < 23.685 ⇒ χ2 ∈ / RR. H0 is not rejected. Thus, there is not
sufficient evidence to support intructor’s claim.

6
4. Given X1 , X2 , ... , Xn ∼ N(µ, σ 2 ).

(4.1) Show that X̄ is an unbiased estimator of µ.

n
! n n
1X 1X 1X nµ
E(X̄) = E Xi = E(Xi ) = µ= =µ
n i=1 n i=1 n i=1 n

Thus, X̄ is an unbiased estimator of µ.

n
1 X nσ 2 σ 2
(4.2) Show that V(X̄) = 2 V(Xi ) = 2 =
n i=1 n n

n
! n n n
X X X X
V(X̂) = V ai Xi = a2i V(Xi ) = a2i σ 2 =σ 2
a2i
i=1 i=1 i=1 i=1

n
X
Since ai = 1, then by Cauchy-Swarz inequality, we have
i=1

1 = (a1 · 1 + a2 · 2 + · · · + an · 1)2 ≤ (a21 + · · · + a2n )(12 + · · · + 12 )

n X σ2 n
X 1
So, a2i ≥ ⇒ σ2 a2i .
i=1
n i=1
n
Thus, V(X̂) ≥ V(X̄).