Viterbi: dynamic programming for HMM

Recursion:

vl(i+1) =

el(xi+1) max(vk(i)akl)

si =

Most probable path

k=2

states

Previous slide

Next slide

Back to first slide

View graphic version