Review

Abstract

In response to the explosive growth of the Internet's information volume, how to discover and track changes in social media has gradually become a key issue in the field of data mining. Especially, how to find the popular and hot topics has become an important issue in social network analysis. In this paper, we review the topic models applied into the social network analysis. At this stage, this kind of problem is mainly solved by the method based on evolutionary clustering. The method can be subdivided into two methods: the topic model and the matrix/tensor decomposition model. This article focuses on the analysis of a class of the most important Latent Dirichlet Allocation (LDA) model from the topic model, as well as a variety of topic models derived from the LDA model, and a brief discussion of future research directions.