Re: some hidden markov understanding questions...
- From: "Ross Clement (Email address invalid - do not use)" <clemenr@xxxxxxxxxx>
- Date: 1 Mar 2006 02:06:55 -0800
Hopefully my suggestion was good, but I garbled the description of what
I did. The inside-outside algorithm is used for parsing text according
to PCFGs, not training the parser. And what's worse is that I can't
remember the name of the algorithm I used for training the parser
(fixed grammar, learning of rule probabilities only) except that it was
a form of expectation-maximisation and that I used the description
found in _Foundations of Statistical Natural Language Processing_ by
Christopher D. Manning, & Hinrich Schutze which I do not have to hand.
Not that anyone asked for this information, but I thought that I would
put in the correction.
Also, I'll point out that the parallel "worked example" and program
approach to debugging is often very hard work, and I try to avoid it if
possible. However, when writing programs for things of similar
complexity to HMMs, I end up using it more often than not :-(
Cheers,
Ross-c
.
- References:
- Re: some hidden markov understanding questions...
- From: John Uebersax
- Re: some hidden markov understanding questions...
- Prev by Date: Re: some hidden markov understanding questions...
- Next by Date: Re: some hidden markov understanding questions...
- Previous by thread: Re: some hidden markov understanding questions...
- Next by thread: Re: some hidden markov understanding questions...
- Index(es):
Relevant Pages
|