Study On The XML-based Web Usage Patterns Discovery Model

Coupled with the flourishing development of The World Wide Web, Due to the larger and more complex scale of the web sites, some central work such as the design of the sites, the web services and e-commerce has become more and more complex. Useful knowledge is supposed to be found out for enterprises and sites designers from a huge amount of information. Dealing with the data of sites using a new effective web usage mining solution has been a hospot in the field of data mining internationaly.Traditonal universal data mining solution, such as mining technology based on web server log is implemented in some situation. The user's navigate path is recorded in the web server log when accessing the web site, by which both the authenticity and the integrality of the dataset is unbelievable. A new web usage patterns discovery model using XML is promoted which could be considered as a new way to optimize the traditional one.The data from client, the server log and databases got together dynamicly, the data source become more abundant and accurate after integration, preparing well for the next step of data mining. In order to optimize the web mining model, some technology relative with XML is needed. According to session schema defined previously, data is recorded in the XML documents, making it simpler and more accurate for users and sessions to be identified. Different structerd data transformed into XML format may transfer through different platforms and be stored in various databases. Not only single-dimensioned but also multi-dimensioned dataset could be created and traditional mining arithmetic like FP-tree growth can be optimized. The accurance and the efficience of the system are enhanced entirely. Then, a new web usage mining model is promoted.In order to implement the XML-based web usage patterns discovery, the data predispose and patterns mining need to be implemented. The former is supposed to be developed on .NET platform and the latter is implemented using the method of association rule mining, when mining frequent pattern, a new FP arithmetic based on XML is applied.