This preview shows page 1. Sign up to view the full content.
Unformatted text preview: w distribution distribution parameters parameter estimate (from m training samples) w w w w x x n , N 2 , x = = m x m 1 i i = n n s s ( ) 2 2 1 s n ~ 2 1 n c s 4 = = m s c 1 m 1 i i 4 = c 4 c 1 2 4 c 4 c 1 2 4 R R complicated d R 2 = = m R d 1 m 1 i i 2 = d 2 d 3 d 2 d 3 np D Bin(n,p) p m p p m i i = = 1 np ) p 1 ( np p n ) p 1 ( p n p n D p = p n ~Bin(np) p m p p m 1 i i = = p n ) p 1 ( p p n ) p 1 ( p c x Pois(c) c m x c m 1 i i = = c c c c u n x u = x~Pois(n ) m u u m i i = = = 1 n u n u use distribution parameter estimates in expressions for w and w...
View Full
Document
 Spring '10
 M

Click to edit the document details