Performance evaluation of topic modeling algorithms for text classification

Abstract

Text Classification is a paramount task in natural language processing. Topic modeling algorithms have been used with a lot of success for text classification. We evaluate different topic modeling algorithms for two tasks: (1) Short text or sentence classification and (2) Large text or document classification. We give an extensive performance evaluation with the help of a wide range of performance metrics for three topic modeling algorithms on both of these tasks using three publicly available datasets. ©2019 IEEE.