# Forward and backward algorithm for HMM

I am looking for example based explanation for HMM's. How is that the HMM has P(O|HMM model) will have 2T* N^T (T is the length of observation, N is number of hidden states.) And how this is solved by forward and backward algorithm. I need some example. - Thanks in Advance

-
How is what solved ? $2T*n^T$ what? –  Peter Flom Oct 2 '12 at 21:20
The question is asking about the forward backward algorithm which is used to do MAP inference in hidden markov models, and how it can explore the exponential in T number of possibilities efficiently. en.wikipedia.org/wiki/Forward%E2%80%93backward_algorithm –  rrenaud Oct 2 '12 at 21:40