基于主观贝叶斯的点击流数据分析应用研究  

Application and Research of Clickstream Data Analysis Based on Bayes

作　　者：王军豪[1] 彭岩[1] 

机构地区：[1]首都师范大学信息工程学院,北京100037

出　　处：《计算机技术与发展》2008年第7期116-118,121,共4页Computer Technology and Development

基　　金：北京市优秀人才培养资助项目(20061D0501600220);北京市强教计划资助项目

摘　　要：通过对不确定性推理和主观Bayes方法的分析研究,提出将主观贝叶斯方法应用到点击流数据分析系统中。在用主观贝叶斯方法进行推理计算时,针对Web日志文件中记录信息的不完备情况,应用了证据的不确定性推理,在系统中对用主观Bayes方法得出结论进行专家分析评估,来确定用户对网站内容的关注程度和上网喜好,从而掌握网站对用户的黏着度,进而为优化网站提供依据,为进一步建设更有吸引力的网站提供决策支持。Through analysis and research of uncertainty reasoning and Bayes, presents a method that applied subjective Bayes approach in click stream data analysis system. In view of the incomplete information recorded in the Web log, while using subjective Bayes methods of reasoning, it has applied the evidence of the uncertainty to carry on expert evaluations with the conclusions drawn from the subjective Bayes method. It can help to master the degree of attention of the user for the site content and their preferences and interests; also, it can grasp the website＇s tenacity degree to users, then provide more personalized Web services support for optimizing the web.site, and provide decision support for further construetious with more attraction.

关 键 词：不确定性 贝叶斯 点击流 概率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]