An Overview of Hierarchical Topic Modeling

Topic correlations are common in real-world textual information. However, classic topic modeling isn't able to model the correlations among topics because of a single distribution over topics in each document. Therefore, hierarchical topic modeling is designed to relax this restriction. In this paper, hierarchical topic modeling is summarized by analysis of existing studies, especially, two important representatives of hierarchical topic models and their extension are focused on. To the best of our knowledge, this is the first effort to review the development of hierarchical topic modeling.