Dr. Hackney STA Solutions pg 117

# Dr. Hackney STA Solutions pg 117 - 7-20Solutions Manual for...

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 7-20Solutions Manual for Statistical Inferenceb. BecausefX(x) is in the exponential family,∑iXiis a complete sufficient statistic andE (nX(1)|∑iXi) is the best unbiased estimator ofλ. Because E (∑iXi) =nλ, we musthave E (nX(1)|∑iXi) =∑iXi/nby completeness. Of course, any function of∑iXithatis an unbiased estimator ofλis the best unbiased estimator ofλ. Thus, we know directlythat because E(∑iXi) =nλ,∑iXi/nis the best unbiased estimator ofλ.c. From part (a),ˆλ= 601.2 and from part (b)ˆλ= 128.8. Maybe the exponential model is nota good assumption.7.50 a. E(a¯X+ (1-a)cS) =aE¯X+ (1-a)E(cS) =aθ+ (1-a)θ=θ. Soa¯X+ (1-a)cSis anunbiased estimator ofθ.b. Because¯XandS2are independent for this normal model, Var(a¯X+(1-a)cS) =a2V1+(1-a)2V2, whereV1= Var¯X=θ2/nandV2= Var(cS) =c2ES2-θ2=c2θ2-θ2= (c2-1)θ2.Use calculus to show that this quadratic function ofais minimized ata=V2V1+V2=(c2-1)θ2((1/n) +c2-1)θ2=(c2-1)((1/n) +c2-1)....
View Full Document

## This note was uploaded on 02/03/2012 for the course STA 1014 taught by Professor Dr.hackney during the Spring '12 term at UNF.

Ask a homework question - tutors are online