Afamilyofdistributions f(y; ) is saidtohave monotone likelihoodratio (MLR) ifastatistic T(y) exists suchthatwhenever < , ()~(

Question:

Afamilyofdistributions f(y; θ) is saidtohave monotone likelihoodratio (MLR) ifastatistic T(y) exists suchthatwhenever θ

′ < θ, ℓ(θ)~ℓ(θ

′) is anondecreasingfunctionof T. Forany suchfamily,themostpowerfultestof H0: θ = θ0 against Ha: θ > θ0 forms P-valuesfromvalues of T at leastaslargeasobserved.

(a) Showthatanydistributionintheexponentialfamily f(y; θ) = B(θ)h(y) exp[Q(θ)R(y)]

(see Exercise4.75)thathas Q(θ) monotone increasingin θ has MLRwith T = Σi R(yi).

(b) Showthatthebinomialdistributionsatisfies(a)with Q(π) = log[π~(1−π)] and T = Σyi.

Fantastic news! We've Found the answer you've been seeking!

Step by Step Answer:

Question Posted: