STA261H1S Term Test 1 (Wed Afternoon)
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
STA261H1S Term Test 1 (Wed Afternoon)
2022
Question 1
(a) [6 Marks] Let’s assume that X1, X2 , ..., Xn f (x; θ), where the density is given by,
f (x; θ) =
when, 0 < xi < θ and f (x; θ) = 0 otherwise. Let’s assume we have the following estimator of θ θˆ = 2
Find the Mean Squared Error (MSE) of θˆ.
Solution
Here we can see that E(θˆ) = E(2) = × ∑ E(Xi).
It can be shown that E(Xi) = . Thus, E(θˆ) = n = θ . Thus the bias E(θˆ) − θ = 0 The
Var(θˆ) = Var(2)
= Var(Xi);
4n θ2
=
n2 12
2
=
3n
The MSE is calculated by,
MSE(θˆ) = (E(θˆ) − θ = 0)2 + Var(θˆ) =
(b) [4 Marks] Using the information in part (a) above, is T = θˆ2 = 42 an unbiased estimator of ψ(θ) = θ 2 ? If yes, prove your answer. If not find the bias of T (i.e. Find Biasθ(T)) as an estimator of ψ(θ) = θ 2 .
Solution
E(T) = E(θˆ2 ) = (E(θˆ))2 + V (θˆ) = θ 2 + θ 2 and so T is not an unbiased estimator of ψ(θ) = θ 2 .
Biasθ(T) = E(T) − ψ(θ) = (θ 2 + ) − θ 2 = ■
Question 2
Let’s assume that X1, X2 , ..., Xn ∼ f (x; µ, σ), where the density is given by,
f (x; µ; σ) = exp (− (ln(x) − µ)2 ) , 0 < x < ∞
(a) [4 Marks] Find the maximum likelihood estimator for µ
Solution: The likelihood function,
L(µ | s) = exp (− (ln(xi) − µ)2 )
ℓ(µ | s) ∝ − (∑ (ln(xi) − µ)2
ℓ′ (µ) = ∑ (ln(xi) − µ) = 0
⇒ = ∑ ln(xi)
(b) [6 Marks] Assume σ = 1. Let τ be the 95th percentile (i.e., f (x; µ, σ)dx = 0.95). Find the value of τ as a function of µ . That is find τ = ψ(µ). Using the relationship ψ(.), find the MLE of τ .
Solution: Here, X ∼ lognormal(µ = σ = 1). Thus, ln(X) ∼ N(µ, 1). Then,
P (X < τ ) = 0.95
⇒P (ln(X) < ln(τ )) = 0.95
⇒P (ln(X) − µ < ln(τ ) − µ) = 0.95
⇒P (Z < ln(τ ) − µ) = 0.95
This implies, that ln(τ ) − µ = 1.645 ⇒ τ = exp(µ + 1.645). Using the invariance property of
Question 3
Suppose we have a finite population Π and a measurement X : Π → {0, 1}, where |Π| = 15 and |{π : X(π) = 0}| = 4.
(a) [2 Marks] Determine fX (0) and fX (1).
(b) [4 Marks] For an i.i.d. sample of size 5 (this means sampling with replacement), determine the probability that 5fˆX (1) = 2.
Solution
Note that 5fˆX (1) is the number of 1’s in the sample of size 5 and under iid sampling 5fˆX (1) ∼
Bin(n = 5, θ = 11/15) and so P (5fˆX (1) = 2) = (2(5)) ( )2 (1 − )3 ■
(c) [4 Marks] For a simple random sample of size 5 (without replacement), determine the probability that 5fˆX (1) = 2.
Solution
Let X be the number the number of elements with measurement 1 in a sample of size 5. Then, P (5fˆX (1) = 2) ⇒ P (fˆX (1) = 2/5). That is we need to find P (X = 2) with replacement. Implying hypergeometric rule we get,
( 2(11))(3(4))
( 5(15))
■
Question 4
A random variable X has the following distribution with parameter θ > 0, and the PDF is as follows,
f (x; θ) = exp (− ) ; x > 0
(a) [5 Marks] If we are interested in making inference about the median of the distribution then determine ψ(θ)
Hint: The median M is calculated as f (x; θ)dx = 0 .5. Then solve for M
Solution: Calculating the CDF,
F (x; θ) = exp (− )
= z exp(−z)dz; [substitution]
= 1 − exp( −z) = 1 − exp (− )
Let the median be M . Thus,
1 − exp (− ) = 0.5
⇒ − = ln(0.5) = − ln(2)
⇒M2 = 2θ2 ln(2)
⇒M = θ √2 ln(2)
The median is θ√2 ln(2).
(b) [5 Marks] If we are interested in making inference about the mode of the distribution then determine ψ(θ). You don’t need to perform the second derivative test.
Hint: The mode of X is calculated as by differentiating the PDF (or ln of the PDF) with respect to X and then find the maxima.
Solution: Take the ln of the PDF,
ln(f (x; θ)) = ln(x) − ln(θ2 ) −
∂ ln(f (x; θ)) 1 2x
∂x x 2θ2
1 x
=
x θ 2
x2 = θ 2 x = θ;
The mode is θ .
since,x > 0
Question 5
(a) [4 Marks] Let’s assume X1, X2 , ..., Xn Binomial(m, θ), where m ∈ N is a constant. Using the factorization theorem calculate a sufficient statistic for θ .
So(Hi)lu(nt)t(:)he(M)Fliokfelih(B) θxwrt(1)t
L(θ) = (xi(m))θxi(1 − θ)m−xi
= (xi(m))θ ∑ xi(1 − θ)nm−∑ xi
= (xi(m)) ( )∑xi (1 − θ)nm
, xn), θ) = ( )∑ xi , i.e.,
(b) [6 Marks] Let’s assume X1, X2 , ..., Xn f (x; θ), where the density is given by,
f (x; θ) = θ2θ x − (θ+1)
where, 2 ≤ x ≤ ∞ and θ > 0. Find the maximum likelihood for the parameter θ . Solution: Here the likelihood and the log-likelihood,
n
L(θ) =∏ θ2θ xi(−)(θ+1)
i=1
n
= θn 2nθ ∏ xi(−)(θ+1)
i=1
ℓ(θ) = n ln(θ) + nθ ln(2) − (θ + 1) ln(xi)
= + n ln(2) − ln(xi) = 0
⇒ = ln(xi) − n ln(2)
⇒θˆ = n
The second derivative,
∂2 ℓ(θ) n
= − < 0
Thus, θˆ is the MLE.
Note for TA: Students might use log instead of ln and that is fine. Please don’t penalize.
2022-04-12