THE PRIVATE RECOMMENDATION BASED ON THE ANALYSIS OF USER DYNAMIC BEHAVIOR

The traditional private recommendation system always ignores user dynamic behaviors. In consideration of the problem, the private recommendation based on the analysis of user dynamic behavior is provoked. This method recommends interested users in their virtual communities which are identified in the dynamic behavior network. The dynamic behavior network is built and made up of micro blog hot topics, users and participation behavior relationships. Meanwhile, this method not only considers the short-term dynamic interest, but also takes long-term stability interest into account. In order to get the weighted similarity of interest, establish the long-term interest model and short-term interest model, and trade off their contribution rate. Finally, experiment is done on a micro blog data set, The results show that this method has good effect.