Weighted time sequence text network-based time sequence community and topic detection method

The invention discloses a weighted time sequence text network-based time sequence community and topic detection method. The method comprises the steps of constructing a weighted time sequence text network based on original data; for the weighted time sequence text network, building a topic model-based generation model; constructing an inference process of the generation model by utilizing a Gibbs sampling method; according to the inference process of the model, training the weighted time sequence text network, and extracting community information, topic information, a corresponding relationship between a community and a topic, and a characteristic of change of influence and participation degree of a user in the community along with time; and according to the extracted information, predicting a user behavior. According to the method, time information and weight information in the time sequence text network are subjected to brand-new modeling; the time information of edges in the network is considered and subje