On 01/21/2014 06:30 AM, Nguyễn Hữu Thiện wrote:
Dear all,
Does the implementation of MaxEnt in OpenNLP include any sorts of
regularization (L1, L2). If not, is it possible to include it into the
current implementation?
No it doesn't. The maxent stuff is a bit outdated, even still doing GIS
training.
We would be really happy to have somebody working on it.
A while back we got a contribution for L-BFGS parameter estimation, but
it still
has some issues. In case you are familiar with it, it would be nice if
you could
have a look there too.
Jörn
