Robust PLSA performs better than LDA

Abstract

In this paper we introduce a generalized learning algorithm for probabilistic topic models (PTM). Many known and new algorithms for PLSA, LDA, and SWB models can be obtained as its special cases by choosing a subset of the following "options": regularization, sampling, update frequency, sparsing and robustness. We show that a robust topic model, which distinguishes specific, background and topic terms, doesn't need Dirichlet regularization and provides controllably sparse solution. © 2013 Springer-Verlag.