solpracfinal

# solpracfinal - 7 x x ~ Hbv|{k | t x b}\$ 5m t x |{zyxswA5m t...

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 x x ~ Hbv|{k | t# x b}\$ 5m t# x @ |{zyxswA5m t# vom t# @ vosuE wostk @ m @ m # womsr pwom @ @ g womnlx#Gq ( g wonlxkj` % ihg f @ p @ m # e d # e1C . Then, . # # e'8ex6 9 6 9 i e'I ( i # RAp@ B UyGq 6 B s r 7 9 i e5 ( i # F 9 i Bu ( A!yxq B u B @ s r F eGB B HGEf! F @CB @ u ( A!yxq u B @ s r F eGB B B HFGEf! @CB @ u ( A!yxq u wA!@ v8thq B u s r B @ s r SQ 9 i HhgfpA TRP F @CB @ i F @ a Y B @ a Y cG b`X ! b`X W eRd SQ TRP F @CB @ IHGED!A 9 SQ P F @CB @ Hhgf! F @CB @ VGEU! , conclude that . Hence, (1) The cdf is 7 4 6 8' 4 5" & & '" # #\$" ! # 2301 ()% % and hence Since, so that (2) We know that (3) Let Solutions to Practice Final Exam 1 Q V x C x . x y B C!@ PB r H |Ixx @ @ 9 8 A7Q 7 FD B GECa % B & 7!@ (B Y 6 yx 5% 1 6 3 x 4|v 1 2\$ x B {vyp@ x \$ )p@ (B 'Y B & x | B o@ x y!@ p B "B @ bx #! E| ~ tkcbv ~ W x # x f f t# 7 |x x | f # i g i s r f tj f u % % v zI % 8thq 7 % I f b`X aY 8txq i s r f 7 t En@ @ k b t# ~ (3b) The estimated standard error of is . Now use the delta method: and the estimated standard error of is . The approximate confidence interval is f e# Gbx x E @ |x bx I f x x bv s|{k |f x @ k | tjIn@ Ibx d f ~ # Now, (4) Note that is 0 if 0 if this happens for any . So . This is the same as . So, (3c) The variance of (3a) The mle is is or . The likelihood will be is non-zero only if for all . On the other hand, if |x 8e# {bx 0p "B @ x B o@ x { ! % ! (4c) First, x B p@ (4b) . For , 2 x m W T R USQ Q x { B !@ . The ARE is . E I . B B P B B B ` B B y( B B B B B y( B B `gAB v B B Q u E hg E E P i g "HG E g g B B B AB B n E i . D AB E F S ( B v B i F Y D E # #tqC . # 7tk ( bY a e A B g e # 7tk6C & # IEg 58e1C 3 4 7 # 2 q # ) 1 ! !Q 0(zI & ) ' g y0 % # #tk %I \$" n ( !Q wG i i Rh B !@ B i . . V x ` 0 1 q # % 0 Q 1 1 h Q C 7 1 0 I x Qx Q ||x " ! Q B Cp@ x B 0 B B B % Q B !@ Q kom om Q B )p@ Y B & m The density is for T R USQ (5a) (5b) Let ! 2# # \$ q !Q "hi 1 (5c) (5d) (6a) Let D 9 @8 (6b) as long as . Setting i.e. as long as . Setting where yields . The Fisher information is 3 for some and gives . B Q [email protected] m %t# 6 Q % B!@ &D B ost# 6 % m Q % B!@ % B om Q &D x om , V 'x C x x x x @ % 'x x 1 %5 % @ 7 @ FD B GECa @ 9 8 A7Q x # \$"! x 3 1 @ 1 yx p @ % @ yx % @ x @ hg i s I @ v g @ s i s i @ @ 7 # 6 for . | y0 ( % % w i m 7% 0 # x i m wh # tj| y0 % s ! A8 i @ s 0 % 0 0 q ( % t# # @ # kom ( vod i m i # kom ( # 6 wh # e# @ m oI vozI d vo i m i m 9 @8 g E E B B B B 8 B B i B B 8gAB vthq B B IuE `X i s r aY then and then . B B 8 i For (7a) First, (6c) The variance is . If , . For . A8s # So, ht y0 and % | y0 s for i hg (7b) Write (8a) Let unless (7c) Given . Hence, and and , where be the smallest and largest values. The likelihood is 0 i.e. . Thus, % @ 1 @ This is a decreasing function so x |v 1 2\$ (8b) First, . For . is a point mass at 4 W Q T R USQ . , . So s EC Ri T1H ek ( #C # i 1 7 % 0 1 0 1 0 # 7 0 @ W 0 0 f 0 @ W b @ W W 5 W e# & 7 Q Q 1 1 0 I I i m om Q V x C 3x 7 x x % 0 1 E| d cbb I 1 Q 1Q ! x x ' t# ||x " B x %t# x B %t# 5 6 The density is (9) Parts (i)-(iii) should be straightforward. For part (iv), note that . Let . Then, if is a nonnegative integer, i Q Q 7 Q & x t4 Q Q Q 5 Q 5 6 Q Q x g|b d c| j` x for T R USQ This is Gamma . So the posterior mean is B !@ B ji I f Q % For part (v), note that Q F F )GD Q (8c) For . So (10) Let the density is . The posterior density is . The mean of a Gamma ( . 5 E . The confidence interval is . . The mean is ) is (in this version) . Also, @ ! B !@ B ...
View Full Document

## This note was uploaded on 05/25/2008 for the course STAT 36625 taught by Professor Larrywasserman during the Fall '01 term at Carnegie Mellon.

Ask a homework question - tutors are online