STAT 201B Lecture Notes - Lecture 23: Rejection Sampling, Point Estimation
Document Summary
- - xml or fix to) likelihood lce ) = ( xi - - - xn ) f- col x" )= f- cx. " to > fco) f-cx ) f- cx ) = foe, fcxyo) f- co ) do. E-- eca xn ) credible interval (cid:8869) point estimation. Conjugate pairs : posterior dist same form prior dist " generally fcxio)fco ) t f-co ix ) if cxlohco) takes known form. I - - - ob n f co l x" ) Idea : know how sample dist " tco) want dist " with density glo ) easy to sample pco ) f is similar to geo ) easy sample plot. F is similar to g cos find c > 0 goose of co) to cand npco, U - unit co , i ) cand) Cand otherwise reject repeat above until i get ab. Skooldo normalizing constant a - - ob (final sample ) ago )