Search

It is a hot research topic to obtain user interest by analyzing user's UCC content. Identifying user interest changes in social networks is of great significance for providing high-quality personalized network services. For the cold - start' users with poor UGC, LDA topic model is not effective to obtain their interest, and it is more difficult to identify the interest changes of them. To solve this problem, LDA is firstly used to obtain network top ices, and by which the topic-tag model is constructed. Furthermore, the spatial vector similarity calculation method based on tag association is used to construct the user-topic model. Combined with the learning model, a user interest change recognition algorithm is proposed. The proposed algorithm is applied to the large-scale data set collected by the microblog website, and the network topic, the core users of the network topic and their probability distribution on the interest topic are obtained more accurately, thus the identification problem of user interest change in the model is solved. [ABSTRACT FROM AUTHOR]