一种基于Web资源相似性的序列模式聚类方法

A method of sequence pattern cluster based on Web resources similarity

在Web行为挖掘中,序列模式聚类是一个很重要的课题,其首要问题就是web序列模式间的相似性度量。以往的多数方法都仅仅针对序列本身进行度量,而忽略了系统中资源本身所存在的关联关系以及用户对资源访问的时间因素。针对该问题,提出了一种基于考虑资源相似性的Web访问序列模式的相似度量方法,并且考虑了用户访问资源的时间因素。经过检验,证明能够有效真实地反映实际情况。

In the excavation of Web behavior, sequence pattern cluster is a very important topic, the most important question is the similar measurement between Web sequence pattem. The former most methods merely aim to carry on the measurement at the sequence itself. But it has neglected the incidence relation which resources itself exist in the system , as well as time factor that the user visits to the resources. In view of this question, this article proposeity one kind similar measurement method based on the Web visit sequence pattern that has considered resources similarity, and has considered time factor that the user visit resources. After the examination, the certificate can reflect the actual situation really effectively.