Research And Implementation Of The Portal Web-based Mining Model Of Personalized Service

With the development of enterprises' information, information systems are already widely applied in enterprises. When business functions are being performed, the independent of these systems bring various problems to the management and using of these systems. So a solution that can provide management and searching, and get valuable information is needed. Under this circumstance, the essential of Portal had evolved to today's Enterprise Information Portal from original Portal Website. Enterprise Information Portal helps users to obtain specific data and services without entering systems again. In addition, Portal, as a framework of integrating resources and also an information resource with intensive data, shows more information in a single window, and the structure of the information shows in it become more and more complicated. How will enable users to have a more efficient and effective way to obtain the information they were interested. Personalized services to meet the needs of users is the further development of the Portal direction, is also an inevitable requirement of enterprise information development.Currently, personalized service that Portal provides is based on user's customization, is a coarse-granularity service. On the basis of analyzing Portal's relevant technology, as well as the content and research of personalized services of Portal, and research achievements of adaptive websites, this thesis proposes a method of using the web mining technology to analyze users' visiting data in portal servers so as to get users' visiting models, based on which Portal can provide high-level personalized services. The main research work as follows:Considering Portal's own characteristics and the implementation process of web mining, this thesis proposes system architecture of Portal's personalized services based on web mining.Designs the data reprocessing algorithm and the relevant implementation steps by characteristics of user access data generated by Portal's system. Researching the existing web mining algorithms to improve the traditional matrix mining algorithm. Considering user's browsing interest, give based on the user's browsing interest User-URL matrix algorithm for mining. It has the advantage of being able to more accurately reflect the user's browsing situation.Explains the model system to provide the content and the way of personalized services.Finally, we implemented the model system by our works presented in the thesis, and achieving the target of providing personalized services for users. Completes preliminary analysis and tests about the results of personalized services.