The analysis of QQ group topic based on the LDA
Jin Qiu, Xiaoqiang Di, Dali Yin, Lin Bi · 2016
In the group of instant messaging software like Tencent QQ, there are too many messages and it is hard to quickly grasp the useful messages concerned with the members. In order to solve this problem, we collect more than 300,000 messages from four active QQ groups for three months, and analyze the hot topics by LDA (Latent Dirichlet Allocation) model. At the same time, the topic probabilities are estimated by Gibbs Sampling algorithm, so it is convenient for members to quickly understand the useful messages in the QQ group.