Latent Dirichlet Allocation

A library for the LDA topic modelling algorithm in Python and C.

 __    ____  _____ 
|  |  |    \|  _  |
|  |__|  |  |     |
|_____|____/|__|__|

Usage

The best way to use liblda is through the command line:

./run.py --docs docs.txt --numT 40 --vocab vocab.txt --seed 3 --iter 400 --alpha 0.1 --beta 0.01 --save_probs --print_topics 10

where:

• docs.txt contains one document per line,
• vocab.txt contains the vocabulary (one word per line)
• --save_probs indicates that you want to output the probs phi and theta

Installation

Place the directory liblda somewhere in your Python path.

Features

We have implemented the Gibbs sampling approach which is fairly efficient when done in C. All the rest of the functionality is done in Python so it is very hackable.

Requirements

• numpy (for arrays)
• scipy (for weave)

Project status

The code base works, but is a bit of a mess right now. A rewrite has begun -- in cython.

Author

Ivan Savov, first dot last at gmail