用户Web日志序列模式挖掘研究

Research on Mining Users' Web Log Sequence Patterns

单个用户访问网站能够留下大量的访问信息,合理地挖掘这些信息便能够得到用户个人的访问模式。文中将序列模式挖掘运用到单一用户Web日志上,最终可以得到单一用户的访问序列模式。在序列模式挖掘过程中,将传统的序列模式挖掘概念进行了扩充,对应到单一用户的序列模式上;运用Session来划分时间段,增强了时间的概念;运用概念格的理论,很好地实现了增量序列模式挖掘。并使用一个新的算法,解决MFP(最大前向路径)在Web日志中获取存在的一些问题。

When accessing websites, single user usually left a lot of access information, which can be mined reasonably to acquire accessing patterns of single user. The paper introduces the applying of sequential patterns mining to Web logs of single user, which can eventually discover accessing sequential patterns of single user. During mining sequential patterns, the traditional Concept of sequential patterns mining is extended to correspond to sequential patterns of single user;Session is applied to partition time slice, which strengthens the time concept;Concept lattice is adopted to realize the incremental sequential patterns mining. Besides,it uses a new algorithm to solve the problems in acquiring MFP(maximum forward path) from Web logs.