关于提取Web用户浏览行为特征的研究

当前,Web日志挖掘技术已成为实现网站个性化服务的研究热点.运用Markov模型来预测用户的浏览模式,从而提高站点访问率、为站点重组提供有利信息是该领域广泛采用的方法之一.但传统方法建立的Markov模型,存在着数据冗余复杂、模型庞大繁琐等问题.针对这些问题,介绍了一种改进的Markov模型.其方法主要是在原有模型的基础之上,在数据清洗、用户会话识别过程中删除一些不予考虑的因素,大大简化了建立的Markov模型,提高了Web日志挖掘的效率.

Research on extracting patterns from web user behavior

It is a popular research on web log mining to achieve web personalization. In this field, the user web navigation patterns are usually modeled as a Markov chain to increase web site click rates and provide useful information in reorganizing site. But the original model remains the more complex and redundant data which result in an anfractuous model, It seriousely affects farther research on user navigation patterns. An improved Markov model is introduced. Some irrespective problems are omitted in the process of the data cleaning and the session identification, such as the defeated pageview, the timeouts record and so on. By this way, the model is more simple and more effective than the original.