基于深度优先序列模式挖掘的预取模型  

Prefetching model based on depth-first sequential pattern mining

作　　者：卫琳[1] 石磊[2] 

机构地区：[1]郑州大学升达经贸管理学院资讯系,郑州451191 [2]郑州大学信息工程学院计算机系,郑州450052

出　　处：《计算机工程与应用》2007年第20期169-172,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(the National Natural Science Foundation of China under Grant No.60472044);河南省高等学校信息网络重点实验室开放基金

摘　　要：序列模式挖掘能够发现隐含在Web日志中的用户的访问规律,可以被用来在Web预取模型中预测即将访问的Web对象。目前大多数序列模式挖掘是基于Apriori的宽度优先算法。提出了基于位图深度优先挖掘算法,采用基于字典树数据结构的深度优先策略,同时采用位图保存和计算各序列的支持度,能够较迅速地挖掘出频繁序列。将该序列模式挖掘算法应用于Web预取模型中,在预取缓存一体化的条件下实验表明具有较好的性能。Sequential pattern mining can find the user access patterns hidden in the Web log,which can be used to predict the upcoming Web access objects in the Web prefetching modeLMost sequential pattern mining algorithms are based on Apriori breadth-first search strategy.A new sequential pattern algorithm based on depth-first search strategy is introduced and the mining mechanism based on lexicographic tree is presented in this paper.The data structure of bitmap is used in order to save and calculate the support of sequences fast.By the use of which,a prefetching model is proposed in integrated Web prefetching and caching environment.The experimental results show that the prefetching model based on depth-first sequential pattern mining can have a good performance.

关 键 词：序列模式 深度优先 WEB缓存 WEB预取 WEB挖掘 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]