Søren Hauberg wrote:
> We have some HMM code in the 'statistics' package. I've never used so I
> don't really know how good it is. If GHMM only supports the standard
> (discrete hidden variables + discrete/Gaussian output) then I think we
> might as well implement it from scratch ourselves to avoid having a
> dependency, and avoid having to write wrappers (which might be a pain to
> maintain). If GHMM does more than the basic stuff then I can see that it
> might be relevant to wrap it.

I think it does make sense to wrap GHMM. Not only is it much more 
feature rich than the three functions in the statistics package, it is 
also much faster.

The hmm functions in the statistics package were written with Matlab 
stats compatibility in mind. There are still two functions missing: 
hmmdecode and hmmtrain. I somehow got stuck in hmmdecode as I could not 
make it to give the same results as the corresponding Matlab function 
(although both are still on my todo list).

Arno

------------------------------------------------------------------------------
Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
easily build your RIAs with Flex Builder, the Eclipse(TM)based development
software that enables intelligent coding and step-through debugging.
Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
_______________________________________________
Octave-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/octave-dev

Reply via email to