基于改进蚁群算法的用户有效浏览兴趣路径挖掘  

Mining User's Valid-interested Browsing Paths Based on Improved ACO

作　　者：肖丹[1] 尹春华[1] 

机构地区：[1]北京信息科技大学信息管理学院,北京100192

出　　处：《计算机与现代化》2013年第12期14-18,共5页Computer and Modernization

基　　金：北京市属高等学校人才强教计划资助项目(PHR201106133);北京市教育委员会社科计划面上项目(SM201110772006)

摘　　要：从Web日志中挖掘用户浏览兴趣路径,对于网站重构和产品推荐等商业用途具有重要意义。传统的挖掘算法一般基于用户访问频度,并不能真正体现用户的兴趣点。访问时间是一个能反应用户浏览兴趣的重要因素,用预设的访问时间阈值剔除无效数据,得出有效访问时间。本文对蚁群算法加以改进,用浏览频度和权值因子作为兴趣函数,有效访问时间因子作为信息素函数,提出有效-兴趣度的概念。对算法进行具体模拟,结果表明,本文提出的挖掘算法比传统的算法更能体现用户的浏览兴趣。Mining user’ s interested browsing paths from Web logs , is of great significance for the website reconstruction , product recommendation and other commercial purposes .The traditional mining algorithm is based on users ’ access frequency , and can-not really reflect the users ’ interests.The access time is an important factor which can reflect the users ’ browse interests;by u-sing the preset access time threshold to cull invalid data , the effective access time can be obtained .This paper improved the ACO, using the browse frequency and weight factor as interest function , making effective access time factor as pheromone func-tion, put forward the concept of the effective-interest degree.In this paper, a specific simulation to the algorithm is presented . According to the result of the experiment , the mining algorithm presented in this paper can better reflect the users ’ browse inter-ests than the traditional algorithm .

关 键 词：蚁群算法 时间因子 有效-兴趣度 WEB日志 浏览兴趣路径 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]