基于Markov与PageRank算法的Web日志仿真器  

Web Access Session Simulator Based on Markov Chain and PageRank Algorithm

作　　者：余智学[1] 林文龙[1] 

机构地区：[1]合肥工业大学电子商务研究所,安徽合肥230009

出　　处：《计算机技术与发展》2008年第5期182-184,187,共4页Computer Technology and Development

基　　金：国家自然科学基金资助项目(70672097)

摘　　要：获取可靠的Web访问会话数据是Web使用挖掘(WUM)的重要前提,而很多时候这种数据不容易得到。据此,采用数学建模的方法,设计并实现了一个Web日志仿真器(SSPM,Session Simulator based on PageRank and Markov)。SSPM用Markov链过程模拟用户访问过程,将用户Web访问过程抽象为Markov链,以PageRank算法计算页面重要度,并以此计算Markov初始状态和转移矩阵,获取用户仿真日志。还介绍了SSPM的验证方法。Reliable Web access sessions are crucial to WUM. Guidcd by mathematic modding metliod,abstracts user Web access process as a Markov chain,and designs a Web access session simulator named SSPM. FageRank is calculated by PageRank algorithm, then with the PageRank, Markov initial state and ambelant Matrix also can be obtained. Since user Web access process is simulated by Markov chain, SSPM can easily acquire simulated Web access sessions. How to validate the simulated Web access sessions is also specified.

关 键 词：WEB使用挖掘 日志仿真器 马尔可夫链 PAGERANK算法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]