摘要
针对用户访问模式(UserAccessPattern)问题,首次提出了一种新的挖掘方法MFP(MaximalFrequentPath),该方法采用两种兴趣度的度量标准(在考虑了用户覆盖面的同时,又考虑了个人贡献因素).MFP方法提高了挖掘算法的精度,增加了挖掘算法的可用性,能挖掘出比其它方法更具有普遍意义的模式.并通过理论推导和实验验证了它的有效性.
As far as the issue on web user access pattern (the issue on traversal paths) is concerned, the thesis first introduces an approach named MFP that uses two kinds of interesting criterions to solve the issue : click ratio and percent of users. Because this approach considers both personal contribution and user' s proportion, the patterns after mining have more implied significance than those of single interesting criterions. At the same time we conduct and prove its availability from theory and practice.
出处
《小型微型计算机系统》
CSCD
北大核心
2006年第5期919-923,共5页
Journal of Chinese Computer Systems
基金
北京市教委科技发展计划项目(KM200510016002)资助