基于用户标签的社区发现方法研究

为充分挖掘标签中潜在的用户信息、提供更好的个性化服务,在分析用户标签使用情况的基础上,提出一种面向用户的社区发现方法。综合考虑标签使用频率和稀疏度,并同时使用二者构造一个邻接矩阵进行聚类,最终完成社区发现,并且论证了该方法的可行性。

Research of community discovery algorithm based on user tags

Existing methods for clustering precision of community discovery algorithm are not quite efficient. In order to provide better clustering precision for community discovery, we proposed a new method based on analyzing information of user tags. This algorithm considers user common relationship based on tags, user common weight based on tags, tags sparsity, impact factor of user similarity based on tags, and other factors, to complete community discovery. Experiments show that the algorithm can provide higher clustering precision.