[mlpack] Fwd: Degenerate cases and other GMM problems

[John Demme]

(missed reply-all a second ago)

That patch is somewhat similar to a mod I tried. Mine helped sometimes, but
not in all situations. Both problems still occur with r14883:

$ ./gmm -i gmm_obs0.csv -g 5
[DEBUG] Compiled with debugging symbols.
[DEBUG] Iterations: 42
[DEBUG] EMFit::Estimate(): initial clustering log-likelihood: 1.4249e+06

error: inv(): matrix appears to be singular

terminate called after throwing an instance of 'std::runtime_error'
  what():
Aborted (core dumped)

$ ./gmm -i gmm_obs0.csv -g 8
[DEBUG] Compiled with debugging symbols.
[DEBUG] Iterations: 57
[DEBUG] EMFit::Estimate(): initial clustering log-likelihood: -inf
[DEBUG] GMM::Estimate(): Log-likelihood of trial 0 is -nan.
[DEBUG] Iterations: 83
[DEBUG] EMFit::Estimate(): initial clustering log-likelihood: -inf
[DEBUG] GMM::Estimate(): Log-likelihood of trial 1 is -nan.
[DEBUG] Iterations: 59
[DEBUG] EMFit::Estimate(): initial clustering log-likelihood: -inf
[DEBUG] GMM::Estimate(): Log-likelihood of trial 2 is -nan.
^C

~John


On Tue, Apr 9, 2013 at 2:06 PM, Ryan Curtin <gth671b at mail.gatech.edu> wrote:

> On Tue, Apr 09, 2013 at 02:02:24PM -0400, Ryan Curtin wrote:
> > I recently committed a fix which adds some small values to the
> > covariance matrices so that they don't end up being entirely
> > zero-valued.
>
> My mistake.  I hadn't committed this fix until just now (r148883).
>
> --
> Ryan Curtin       | "I am a meat popsicle."
> ryan at igglybob.com |   - Korben Dallas
> _______________________________________________
> mlpack mailing list
> mlpack at cc.gatech.edu
> https://mailman.cc.gatech.edu/mailman/listinfo/mlpack
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cc.gatech.edu/pipermail/mlpack/attachments/20130409/25f022a4/attachment-0003.html>