|
Boost : |
From: Stjepan Rajko (stipe_at_[hidden])
Date: 2007-05-11 14:34:29
On 5/11/07, JD <jean.daniel.michaud_at_[hidden]> wrote:
> Hi,
>
> I would like to know if there is any interest in an implementation of
> the edit distance algorithm
I'd be interested in a general pattern recognition library - something
that covers different flavors of edit distance and different flavors
of hidden markov models. The two are not that distant algorithmically
- see (shameless self-reference):
http://tinyurl.com/2qyv9n
and
http://tinyurl.com/2qv45o
The first paper draws some common ground between dynamic programming
alignment algorithms (such as edit distance) and hidden markov models,
and the second develops a formal model (semantic network model) which
capitalizes on the relationship.
If I have time, I might spend a part of the summer on re-implementing
the general SNM framework - doing it in a way that also covers basic
HMM and DPA functionality might be doable. I'll take a look at your
implementation.
Stjepan
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk