[developers] Re: estimate and [incr tsdb()]

[Yi Zhang]

Hello Erik,

> in general the file name specified for the -variances option should
> contain either a single number that then sets the global variance for
> the priors (this is the set up used in maxent.lisp), or one number per
> line, individually specifying the variance for the prior used for each
> single feature weight (see the note from rob on
> http://www.isi.edu/~graehl/maxent/malouf/README.malouf.update).

Thanks for pointing me to the note.

> in [incr tsdb()], setting the *maxent-variance* variable to NIL
> instead of a number, simply means that estimation is done without
> using a prior (i.e. the -variances option isn't passed to 'estimate'),
> and this should work fine. using a prior on the weights can help
> reduce overfitting, but you can estimate a model without it.

I think I will disable -variance option and hope it won't hurt too much.

> where exactly did you see messages indicating that it was the
> -variances argument that caused problems? (i couldn't seem to find
> anything related to that in the output and buffer files you sent).

I tried to run estimate on terminal with the same options that [incr
tsdb()] used. The output weights file is just fine if i don't use
-variances option. But with -variances, the weights all become '0',
just like the file ".mem.yzhang.11762.mew" i sent earlier.


Best,
Yi