FINAL ASSIGNMENT STA453H1, 2023
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
FINAL ASSIGNMENT STA453H1, 2023
1 . For a sample size n = 8 suppose that X1 , ...,X8 IID X ⇠ N(µ,σ2 ) with µ 2R, σ >0 and, as usual, let 8µ = ⌃i(8)=1 Xi and 8σ2 = ⌃i(8)=1 (Xi−X)2 .
a) For what value of the constant a is the (squared) distance E(aσ−σ)2 minimized?
b) For the parameter ✓ = µ/σ and the statistic ✓ = µ/σ determine the constant b such that Eb✓ = ✓ as well as the variance varb✓ .
2. location-scale uniform: X1 , . . . ,Xn IID X = a+bZ, Z ⇠ unif[0, 1] , (a,b) 2 R⇥R+
a) Determine (a,b) = mle(a,b) . X−a1
individual marginal distributions for each of a, b and U.
c) Verify that (a,b) is sufficient for (a,b) .
d) Determine the distribution of the pivotal quantity
b
.
3 . Suppose that X1 , . . . ,Xn IID X where X ⇠ unif{1, . . . , ✓}, ✓ 2 N. Let T = X(n) = max(X1 , . . . ,Xn) .
a) Determine the distribution function (D .F.), F✓(t) = P✓(X(n) < t), t 2 R.
b) Determine the probability function (p .f.), p✓(t) = P✓(X(n) = t), t 2 R.
c) Demonstrate that T = X(n) is sufficient.
i) by explicit computation of the cond . distn . ii) by using Neymann’s factorization .
d) Demonstrate that T = X(n) is complete.
e) Obtain the uniformly minimum variance unbiased estimator (UMVUE) for ✓ .
4. exponential family X✓ ⇠ dP✓ = exp (t\ ✓ − K(✓))dP , ✓ 2 ⇥ ⇢ Rp
w. ⇥ = { ✓ 2 Rp l Eexp(T\ ✓) < 1 .
Prove
a) 0 2 ⇥
b) ⇥ is convex
c) E✓T = DK(✓), var✓T = D2 K(✓)
d) T is sufficient for ✓ 2 ⇥
e) T is complete for ✓ 2 ⇥ .
5. variable carrier model
(Xn,n 2 N) IID X ⇠ unif(C) w. C co ⇢(nv)ex Rp , 0 < vol(C) < 1
a) Prove that at each sample size n, the mle, C = Cn, is the smallest convex set containing the sample points x1 , . . . , xn.
b) Letting d(s,t) =^⌃1(n)(si−ti)2 denote the usual metric on Rp, for any A ⇢ Rp, define d(A,t) = infa2A d(a,t) and verify that ld(A,t)−d(A,s)l < d(s,t)
c) Finally, let 6(C,C) = supt2C d(C,t) and prove that 6(C,C)!(wP)1 0 .
2023-04-18