Research On Technique Of Web Log Mining

With the swift and violent development of Internet in amount, scale and complication, Web became an effective platform on which people communicate and manage. A great deal of data accumulates in the Web. In this case, it turns into a difficult problem how to discover the knowledge we need.In order to solve that problem, Web mining technique emerges, as the times require. The technique of Web log mining is an important part in the research field of Web mining. Data preprocessing plays an essential role in the process of Web log mining. The author probes into the main task of data preprocessing deeply, and introduces a plan. This plan bases on the user's visiting time and lowest visiting time, judges this page whether is user's interested page. Then delete the page that the user is not interested in, seek an ultra-link that can effectively reduce the searching way (reduces searching expense), and re-defines the Web transaction.After data is preprocessed, it is time to select a method such as clustering, classifying and associating regulation to mine the web data according to the given demand. The cluster technique is introduced. The content and methods of Web clustering is analysed. Although data has been processed, the amount of data is still very great, so ISODATA (Iterative Self-Organizing Data Analysis Technique A) method is used. But this method's initial value is sensitive and it is likely to trap into local best value. An improved method is put forward and to validated using the trial data.