个性化服务中用户兴趣聚类算法研究

讨论了个性化服务中用户兴趣建模对聚类算法的要求,指出经典聚类算法应用于用户兴趣聚类时的不足。在基于图论的K近邻聚类算法的基础上进行改进,提出一种基于相似度的聚类算法。实验证明,与K近邻算法相比,该算法能够显著提高聚类质量,有效区分孤立点,适用于用户兴趣聚类。

Research on user interests clustering algorithm in personalized service

This paper discusses the user modeling's requirements to the clustering algorithm in personalized service,then the disadvantages of classic clustering algorithm are given.With the improvement of the K-nibblers clustering algorithm based on the graph theory,the clustering algorithm based on the similarity is presented.The experiment results show that this algorithm can improve the quality of clustering obviously and identify the noises correctly compared to the K-nibbler clustering algorithm.