摘要： 介绍Web日志挖掘预处理过程中一些特殊情况的处理方法，在会话识别阶段给出一种基于过滤框架网页与会话重组相结合的会话识别方法，在会话识别之前进行框架网页的过滤，根据传统的会话识别方法构造初始会话集，使用会话重组算法对初始会话集进行优化。实验结果显示，该方法提高了会话识别的质量。

关键词: Web日志挖掘, 数据预处理, frame页面, 会话识别

Abstract: This paper introduces some special processing methods of preprocessing in the field of Web log mining. The algorithm based on frame page filtering and session identification is proposed at the stage of transaction recognition. Before identification, frame pages are filtered, and initial session sets are constructed by traditional method of session identification. The quality of session set is improved using a method of union and rupture. Experimental results show the algorithm enhances the quality of session identification.

Key words: Web log mining, data preprocessing, frame page, session identification

中图分类号: 

• TP393