This repo contains a customized fork of the MALLET machine learning library in Java (http://mallet.cs.umass.edu).

The project is meant to house new packages (and some refactoring of old packages) to the base library that make it more useful for modern data science-y modeling and such. These plans include:

• active learning algorithms
• new interfaces for dual-supervision classifiers
• a regression package
• better support for elastic-net regression models
• a matrix factorization package
• an improved HMM model
• various other utility classes

One step at a time.