Variational Learning for the Inverted Beta-Liouville Mixture Model and Its Application to Text Categorization

TL;DR

This paper introduces a variational learning approach for the invert Beta-Liouville mixture model, enabling tractable inference and demonstrating effectiveness in text categorization tasks.

Contribution

It develops an extended variational inference framework for IBLMM, allowing analytical solutions and improving modeling of positive data.

Findings

Effective in modeling positive data

Achieves tractable inference for IBLMM

Performs well in text categorization

Abstract

The finite invert Beta-Liouville mixture model (IBLMM) has recently gained some attention due to its positive data modeling capability. Under the conventional variational inference (VI) framework, the analytically tractable solution to the optimization of the variational posterior distribution cannot be obtained, since the variational object function involves evaluation of intractable moments. With the recently proposed extended variational inference (EVI) framework, a new function is proposed to replace the original variational object function in order to avoid intractable moment computation, so that the analytically tractable solution of the IBLMM can be derived in an elegant way. The good performance of the proposed approach is demonstrated by experiments with both synthesized data and a real-world application namely text categorization.