[Biojava-l] DP forward algorithm

[Matthew Pocock]

Hi Zhen,

For both the forwards and backwards algorithms, you
can call them in the same way as viterbi. This will
just give you back the score. You can also call them
with an explicit dynamic programming matrix. In this
case, you get the matrix back with all the values
filled in. Both Forward and Backward give you the
probability that the sequence was generated in some
way (any way) by the model. If you look in the
matricies for forwards & backwards, the cells give the
probability that the model made all of the sequence up
to and including that position (and ending in that
state). For the forwards algorithm the probabilities
are for all the sequence from the beginning to there,
and for backwards it's all the sequence from the end
back to there.

Matthew

 --- "Ren, Zhen" <zren@amylin.com> wrote: 
> Hi, there,
> 
> The three main DP operations are Forwards, Backwards
> and Viterbi. Forwards and Backwards calculate the
> probability of the sequences having been made in any
> way by the model. Viterbi finds the most supported
> way that the sequence could have been made.
> 
> I have seen the Dice example that demonstrates how
> to use Viterbi algorithm.  And I guess it would be
> the similar way to use Forward algorithm.  However,
> it's not like what I thought.  Can someone give me a
> direction how to use the Forward algorithm in DP
> class?  Is any tutorial regarding DP and HMM
> available?  More documentations about DP and HMM
> classes in BioJava other than Javadocs are also very
> helpful.
> 
> Thank you.
> 
> Zhen
> 
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l@biojava.org
> http://biojava.org/mailman/listinfo/biojava-l 

__________________________________________________
Do You Yahoo!?
Everything you'll ever need on one web page
from News and Sport to Email and Music Charts
http://uk.my.yahoo.com