[Python-projects] HMM test cases

[Ludovic.Aubry]

On Mon, Feb 11, 2008 at 11:39:45AM +0200, Orestis Markou wrote:
> Replying to python-projects for posterity. In my case I have ~350 states 
> and ~460 observations and the numpy performance is good enough. I'll look 
> into potential improvements, though.
In our implementation the costly part was training using Baum-Welch
algorithm which I can't see in your implementation.

I just had a quick look but it seems you do training with
observation+state (not having to guess the state), am I right?

from what I remember the C implementation was faster by a factor of 5 to
10.

-- 
Ludovic Aubry                                 LOGILAB, Paris (France).
http://www.logilab.com   http://www.logilab.fr  http://www.logilab.org