Community Discovery Algorithm Based on User Behavior Similarity

With the rapid development of Weibo network, community discovery has become an emerging research hotspot. It is found that community networks help operators understand the network model structure and user characteristics and provide personalized services for users. At present, most researches on Weibo community mining only focus on the network structure and the connection of edge nodes, while ignoring the content generated by users, resulting in a lower accuracy rate of community discovery algorithms in practical applications. This paper comprehensively considers the network structure and community user node content, and proposes a community discovery algorithm based on user behavior similarity. After the text data is preprocessed, the topic feature mining is performed according to the LDA, the user behavior information is extracted, the user topic feature words are extracted from the Weibo text, and the attribute characteristics of the user behavior similarity are increased. Taking this eigenvalue as one of the evaluation indexes of the similarity module function, combining the connection relationship and behavior similarity between users, clustering them based on similarity, finally obtaining the community structure and experimenting on the real data set. The experimental results show that the optimization algorithm has strong adaptability in social network systems, and the community partitioning effect is better.