User Analysis Based on Fuzzy Clustering
In order to solve the problem of user-classification to reflect the features of web users inflexible, a novel user classification model was presented in this paper. By introducing the concept of time discretization and applying fuzzy equivalence relation clustering to classify web users, the model can rationally solve the user classification problems. Empirical results showed that the output of user classification was not unique and the parameter 6 should be adjusted based on applications. Compared to those hard clustering, this model is proved to be more effective to classify web users.
fuzzy clustering method web-logs preprocessing user classification
Ming Yang Hong Li
Department of Information System, BeiHang University, Beijing, 100191, China
国际会议
北京
英文
194-196
2009-07-24（万方平台首次上网日期，不代表论文的发表时间）