A Topic Modeling Toolbox Using Belief Propagation

TL;DR

This paper presents a new toolbox for topic modeling that utilizes belief propagation algorithms, offering an alternative to traditional methods for models like LDA, ATM, RTM, and LaLDA, with implementations in C++/Matlab/Octave.

Contribution

The novelty lies in applying belief propagation algorithms to various topic models, providing a flexible and extendable toolbox for researchers and practitioners.

Findings

Implemented BP algorithms for multiple topic models

Compared with existing packages, offering a new inference method

Source code is freely available for community use

Abstract

Latent Dirichlet allocation (LDA) is an important hierarchical Bayesian model for probabilistic topic modeling, which attracts worldwide interests and touches on many important applications in text mining, computer vision and computational biology. This paper introduces a topic modeling toolbox (TMBP) based on the belief propagation (BP) algorithms. TMBP toolbox is implemented by MEX C++/Matlab/Octave for either Windows 7 or Linux. Compared with existing topic modeling packages, the novelty of this toolbox lies in the BP algorithms for learning LDA-based topic models. The current version includes BP algorithms for latent Dirichlet allocation (LDA), author-topic models (ATM), relational topic models (RTM), and labeled LDA (LaLDA). This toolbox is an ongoing project and more BP-based algorithms for various topic models will be added in the near future. Interested users may also extend BP algorithms for learning more complicated topic models. The source codes are freely available under the GNU General Public Licence, Version 1.0 at https://mloss.org/software/view/399/.