Research On The Technology Of Web Log Mining

With the development of Internet, widely application of WWW, and electronization of customers'behavior, it is possible to collect the data about the user traversal and analyze their action further. Now we are confronted with how to use these complex data to obtain the valuable information and knowledge that we can understand. We can use Web Log Mining to solve the problem.Web Log Mining analyzes and researches system based on data mining uncovers the hidden regulations among the interactive data between a web server and its users, in order to get the frequency and behavior pattern of user accessing the site, so web administrator can perfect the structure of web site and hype-line among pages, improve service of web and performance of web site. Further more Web Log Mining can feedback some abnormal access information to web administrator to enhance security of the site.This paper analyzes and researches the Web Log Mining from the following section detailedly. In the first part we present the significance and background of the research and the current research situation in home and abroad; then we summarize Data Mining, Web Data Mining and Web Log mining, show the relationship between them. Secondly we research data preprocessing technology in Web Log Mining, and analyze all tasks in every phase of traditional data preprocessing technology detailedly; then the paper proposes an algorithm which is based on traditional data preprocessing to simplify the steps in data preprocessing. Experiment indicates that the algorithm can improve the speed without lowing the accuracy of preprocessing. Thirdly the paper introduces several algorithms which are used typically in Data Mining, and study Apriori algorithm in association rules and compare its some improved method , then the paper proposes adopting digitalization to carry out Apriori algorithm. In next part the paper introduces the procedures of Web Log Mining and gives some instances. At last, the paper reviews our researches and proposes conclusions including advantage and disadvantage points.