Classification model of network users based on optimized LDA and entropy

The classification of network users is very important in user behavior analysis. The algorithm which was based entropy and latent Dirichlet allocation (LDA) was used in this paper. It is important but difficult to select an appropriate number of topics for a specific dataset. Entropy was first used to solve the problem. A concept named difference-entropy was built to determine the number of topics. Experiments show that the proposed method can achieve performance matching the best of LDA without manually tuning the number of topics.