STAT 201B Lecture Notes - Lecture 30: Olx, If And Only If

• School: Berkeley
• Department: Statistics
• Course: STAT 201B
• Professor: Elizabeth Purdom

Document Summary

Bayes risk : f- ( 0-7=1 rcpi , f) = fo" rcp , pi ) df - r ( pi , f) =n4cntf) " Bayes risk r cf , e) = fr ce ix hex ) dx. , e) ix ) sq error loss = ecco - e 51 7. E : info ella - e) " ix ) F- cy - c) (cid:15482) c = ey. = eco ix) = tiny b"t yn. For any posterior distr olx minimizing bayes risk with e-eloi x) ab s error posterior median zero - one loss posterior mode. X11 - bin cn , p ) p n beta ( a , b ) Minimax rules t admissibility unique minimax estimator for o (cid:15482) admissible. - xp x - ny cq , ttp ) Xi indef n cfi , t) estimate ai of. Xi mean of nobs yi - - - yn.