Mining System Analysis And Design, And Usage Patterns Based On The Frequency And Preference

Internet has been developing quickly at present, and has been applied to all kinds of industries. More and more companies,organizations and people publish and search information on the Internet. Internet not only is full of text,audio,video and other multimedia information but also includes linking between pages, usage while people surf on the net. People wish to discovery some useful knowledge which hided in the rude information by data mining technology, however, it's hard to do this because of its own features.People suggest to discovery useful knowledge hided in the huge information by Web data mining, combing data mining technology and the features of Internet, to improve the performance of server, personalization user accessing and business intelligence and so on.First, this paper introduces some theory knowledge, such as the background knowledge of Data mining, the conception of Web data mining and the classify of Web data mining, emphasizing on the knowledge of Web usage mining, and analyzing the mining arithmetic in detail, from the definition of arithmetic, execution steps, execution efficiency and the analysis of arithmetic at last.Next, we put forward a framework of usage mining. The framework has three modules; they are data pretreatment, usage mining, patterns analysis and visibility. In the module of data pretreatment we put forward a new kind of data source instead of former server log, not only log user's usage roundly but also execute simple and fast. In the module of usage mining we improve the classical association rules according the features of Internet, so the patterns mined from usage could delegate most of users' frequent and preferred patterns really. In the module of patterns analysis and visibility we introduce some popular technique, emphasizing OLAP and visualizing.At last, we introduce some applications of patterns. The patterns have been applied to personalization user accessing,improve the features of server and business intelligence and so on. However, we suggest applying the patterns to intrusion detection, to detect users' visiting by the patterns mined, so it will help to improve the security of Internet.