How to use custom topics in MALLET?
I'm trying to infer the distribution of topics for a document over a set
of custom topics that I have generated outside of MALLET. I'm wondering
the following:
"mallet -train-topics ... output-model modelfile" outputs a binary file
"modelfile". What I would like to do is to create modelfile myself and use
this to classify new documents. Does anyone know where to find a
description of the format of this file? I have searched online to no
avail. The goal here is to convert the topics I already have into the
binary format MALLET can use, then infer the topic distribution for new
documents.
Is there a better way to accomplish this task? I am completely open to
other software that may be more naturally suited to this task.
Thanks!
No comments:
Post a Comment