ealdent / lda-ruby Goto Github PK

So, my research project partner and I were wondering "can it infer the topic distribution of new (unseen) documents?"

If that's not a current feature, do you think it would be difficult to add? Maybe it's something I can help with.

Thanks,
Jason

Hello,

I've noticed that the best way to create a corpus for us is to use the DirectoryCorpus way because we hardly have the .dat files available ahead of time. However, we are looking to be able to identify the documents inside the corpus object after running the analysis and we noticed that the filename does not get populated for each document.

Is this an issue or am I missing something? Thanks!

At first, thanks for this gem.

By this gem I can create a topics (categories). That is perfect! Is there a way how to classify document by the created topic? What would you suggest?

Hello-

I'm having some trouble with segfaults and lda-ruby. At first I thought it was something to do with the way I was using it, but it turns out that the unit tests themselves segfault, so I'm guessing it's not in my code--- or, at least, not only in my code. :-) Here's the output of running "rake test":

https://gist.github.com/1051716

As you can see, it looks to be segfaulting in the EM code somewhere.

Any ideas?

offending line: lda.em("random")

ouput:
"
[BUG] Segmentation fault
ruby 1.8.7 (2010-01-10 patchlevel 249) [i686-darwin10]

Abort trap
"