Study On Web Usage Mining Based On Fuzzy Clustering

Web Servers are now evolving themselves from Server-Centered to Client-Centered Servers,so the trend of Internet web sites'development is Personalization Intelligent Web sites. Personalization Intelligent Web site should cater to every web user's interests and adapt itself frequently to trace the fluctuation of web user's interests.in more detail,when an customer of a Personalization Intelligent Web site is visiting this site,he should have such a familiar feeling that it seems the whole web site is designed for him and it seems bosom friend because it is very convenience to him,and all the contents are very useful which he is interesting in..With the explosively growth of information sources available on the World Wide Web,it has become increasingly necessary for users to utilize automated tools in find the desired information resources.So the application of data mining techniques to World Wide Web(referred to as Web mine) has been the focus of several recent research projects and papers.At present,the research on the Web Usage Mining mostly concentrated on the user access pattern which statistic clustering techniques are ofen used to analyse user' preference on pages.However,this approach can only sort each user session into a single cluster.That is,it ignores a user session may contain several browsing preference by assuming a user session include only a single preference.According to this insufficiency,fuzzy clustering techniques were proposed instead while in detail,mostly are fuzzy equivalence relationship clustering or graph clustering technique in domestic. At the same time,those methods can only use distance of pages to caculate the similarity between sessions.Therefor,if users browse the same web page by different paths,that causes wrong results.This research proposes a framework which combines the target function fuzzy clustering and association rules.At first,this approach filters out the noisy data (outlier hyperlink pages) which may cause false result.Then,it employs association rules to caculate the confidence of the rule as the association between different URL addresses which replaces the formerly method based on distance.And then,caculate the similarity between the sessions.At last,a fuzzy clustering technique based on target function named FC-MDE is adopted to cluster the user sessions which can sort the sessions which contain closely interests into a cluster and find the frequently accessed pages in a cluster.In the end,we come up with a logic frame of a Personalization intelligent recommendation web site.