Web日志挖掘  

Research on Weblog mining

作　　者：宋爱波[1] 胡孔法[1] 董逸生[1] 

机构地区：[1]东南大学计算机科学与工程系,南京210096

出　　处：《东南大学学报（自然科学版）》2002年第1期15-18,共4页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金资助项目 (79970 0 92 )

摘　　要：提出了一种新颖的MBP算法 ,它利用关联规则挖掘发现的频繁项目集以加快速度 ,能找出所有满足阀值约束的频繁浏览路径 ,该算法是很有效的 .同时 ,针对Web浏览和日志文件固有的模糊性和不确定性 ,还讨论了Web页面的模糊聚类问题 .最后 。Similar customer groups, relevant Web pages, and frequent access paths can be discovered by mining Weblog files. Based on a survey of current Web mining research, an algorithm of MBP is presented. It uses frequent items found during association rules mining. MBP can find all frequent access paths meeting threshold constraint. Experiment shows that MBP algorithm is very effective. At the same time, due to the inherent fuzziness and uncertainty in Web browsing and logging, fuzzy clustering for Web pages is also discussed. Finally a discussion concerning the use of discovered knowledge in recommendation system and adaptive Web site as well as some algorithm methods are also given.

关 键 词：WEB日志 数据挖掘 模糊聚类 推荐系统 自适应WEB站点 MBP算法 规则挖掘 

分 类 号：TP393.092[自动化与计算机技术—计算机应用技术] TP18[自动化与计算机技术—计算机科学与技术]