gateplugin-LearningFramework icon indicating copy to clipboard operation
gateplugin-LearningFramework copied to clipboard

Add support for topic models by wrapping gensim

Open johann-petrak opened this issue 7 years ago • 1 comments
trafficstars

This will need an even simpler "corpus representation" for text (list of tokens) only.

johann-petrak avatar Jun 15 '18 16:06 johann-petrak

A quicker way and easier to use may be to use Mallet's LDA module. This requires a different Mallet Corpus Representation (based on FeatureSequences and no targets) and may not scale as well as using an OOM representation with a GenSim backend, but would not need any additional software installed and automatically port to Windows.

johann-petrak avatar Aug 04 '18 22:08 johann-petrak