Web用户聚类研究

The study of Web-user clustering

Web日志数据的海量增长,要求聚类算法能高效的从海量数据中得到满意的用户聚类.本文提出了一种新的聚类算法,在聚类前,利用用户兴趣度对海量数据集进行约简、减小数据计算规模,然后再对Web用户进行聚类.实验证明这种方法能减小数据规模、提高聚类效率,并得到满意的用户聚类.

Massive growth of web log data demand clustering algorithm to efficiently get the satifetory user clustering from massive amounts of data. This paper proposes a new clustering algorithm, a weights degree of user interest value to reduce the massive amounts of data was used,the size of computing scales before clustering were reduced, then web users were clustered. Experiments show that this method can reduce the data size,improve clustering efficiency and get the satisfactory user clustering.