Research Of Data Mining Technology Based On Web Log

With the rapid development of Internet and Internet users increasing, the current Web-based e-commerce information system has also got an unprecedented development. Effective use of the abundant data resources, the timely and rapid discovery of useful information, and provide information of interest to users, which have become the topics to be studied for the vast number of researchers. And Web usage mining technology was born and developed in this context.Web usage mining, also known as Web log mining. As Web log records the information of user's access to the Web site, Users'browsing habits and models can be found through the analysis to the information of the access. Thus it provides a theoretical possibility to optimize the site structure, as well as adds some personalized services to the process of the e-commerce.To begin with, data mining technology and personalized recommendation technology were highlighted in this paper. On the basis of the analysis and research of the deficiencies in traditional personalized recommendation algorithm and issues on recommended quality, a novel and effective data mining method was proposed. Secondly, this paper details and discusses the general process of data mining based on the Web-log and focus on the process of data preprocessing.In order to verify the superiority of the proposed algorithm, a comparative experiment of mining algorithm was designed in chapter 5. MovieLens dataset and the simulation data were used for self-simulation experiment.The result of mining shows that, in the case of the data extremely sparseness, project-based collaborative filtering recommendation method is effective to improve the recommended quality. At the same time, this paper has also set up a book recommendation site which intuitively reflects the recommended algorithm in the paper.In the last chapter, the writer summarized the contents and the main work of this topic and looked to the research orientation for the future work.