基于改进完全子图模型的关注对象多社区发现研究

摘要

为实现用户和关注对象的多社区划分，针对完全子图模型不能进行多类分类的缺陷，该文引入了阈值划分方法，提出基于改进完全子图模型的社区发现算法。实验表明：与经典数据挖掘算法K-medoids相比，该文算法具有更高的准确性。%A multi-community detection method based on improved complete subgraph model is proposed using threshold division for multi-community division of users and concerned objects, because complete subgraph model cannot divide users and concerned objects based on multi-classification. Experiment result shows that compared with classical data mining algorithm K-medoids,this method is more accurate.