0% found this document useful (0 votes)
2 views6 pages

Advanced Stat Inference Final Set1 Exam

Uploaded by

gkals0323
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
2 views6 pages

Advanced Stat Inference Final Set1 Exam

Uploaded by

gkals0323
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Statistical Inference - Comprehensive Final 1 of 6

Final Exam - Advanced Set 1


Advanced Comprehensive Version

Name: Student ID: Instructions.

Answer all questions. Show all essential derivations. Unless otherwise stated, use level α for tests and
assume regularity conditions when using asymptotic theory.
Statistical Inference - Comprehensive Final 2 of 6

Problem 1
iid
Let X1 , . . . , Xn ∼ Gamma(3, θ), where the shape is 3 and the scale is θ > 0:

x2 x
 
f (x; θ) = 3 exp − , x > 0.
2θ θ

(a) Find a complete sufficient statistic for θ.

(b) Find the MVUE of θ.

(c) Find the MLE of θ and the Fisher information I(θ) in one observation.

(d) Use the CLT and the Delta method to find the asymptotic distribution of log θ̂.
Statistical Inference - Comprehensive Final 3 of 6

Problem 2
iid
Let X1 , . . . , Xn ∼ N (µ, σ 2 ), where both µ and σ 2 are unknown. Consider

H0 : µ = µ0 versus H1 : µ ̸= µ0 .

(a) Derive the likelihood ratio test and show that it is equivalent to the usual two-sided one-sample
t-test.

(b) Write the corresponding asymptotic Wald, score, and likelihood ratio statistics.

(c) Explain why the three asymptotic tests have the same first-order limiting null distribution.
Statistical Inference - Comprehensive Final 4 of 6

Problem 3
iid Pn
Let X1 , . . . , Xn ∼ Bernoulli(p) and let Y = i=1 Xi . Consider the test

H0 : p = p0 versus H1 : p ̸= p0 ,

where 0 < p0 < 1.

(a) Derive the exact likelihood ratio statistic in terms of Y .

(b) Show how the exact rejection region is determined by small values of the likelihood ratio.

(c) Derive the asymptotic Wald, score, and likelihood ratio statistics and identify their limiting null
distributions.

(d) Interpret the score test as the usual one-sample test for a population proportion.
Statistical Inference - Comprehensive Final 5 of 6

Problem 4
iid
Let X1 , . . . , Xn ∼ Exponential(θ) with mean θ > 0:
1
f (x; θ) = e−x/θ , x > 0.
θ
For a fixed number a > 0, define
η(θ) = Pr(X1 > a) = e−a/θ .
θ

(a) Find a complete sufficient statistic for θ.

(b) Use functional invariance of the MLE to find the MLE of η(θ).

(c) Find the MVUE of η(θ).

(d) Use the Delta method to find the asymptotic distribution of the MLE of η(θ).
Statistical Inference - Comprehensive Final 6 of 6

Problem 5
iid
Let X1 , . . . , Xn ∼ Poisson(λ), λ > 0. Consider

H0 : λ = λ0 versus H1 : λ ̸= λ0 .

(a) Derive the score function, the Fisher information, and the MLE.

(b) Derive the Wald statistic, score statistic, and likelihood ratio statistic.

(c) Show that all three statistics converge in distribution to χ21 under H0 .

(d) Explain, using Taylor expansion, why these three tests are asymptotically equivalent.

You might also like