This paper studied the discovery of social topics in social networks using an improved label propagation algorithm.To address the problem of traditional algorithms easily falling into local optima,it selected neighbor nodes during label propa-gation based on the similarity between nodes.To solve the randomness issue in label updates of traditional algorithms,it used the node influence to update labels by incorporating the opinion interaction process from the HK opinion dynamics model.The experimental results show that the proposed method,in the best case(k=0.9),improves stability by 31％and modularity by 78％compared to the original algorithm and outperforms several other improved algorithms.It demonstrates that the proposed algorithm performs better in discovering topic communities in social opinion networks compared to the original algorithm and other improved algorithms.
关键词
标签传播算法/舆情社交网络/HK模型/主题社区发现
Key words
label propagation algorithm/public opinion social network/HK model/topic community discovery
维普
万方数据