Research Of Users Browsing Behavior Based On Path And Web Mining

In the process of interaction between the user and the product (especially web browsing), a large number of behavior data have been feedback. It has been many Internet companies’topics that find out about the users’mental and love by means of mining and analysis surfing data deeply, and then, to improve product and users experience.To study the users browsing behavior, there is an effective method that collects the web log when they are surfing. And it has been a success that study the users browsing behavior according mining the web log. This paper hopes based on this success, combine with the current popular Hadoop platform and data warehouse technology, to make the users browsing behavior which based on web log mining systematize and engineered. So as to make it become a project that can be used directly in the daily production of the enterprise, and support the product development, operation and management of enterprises much better.Based on path and web mining, this paper researched about the users browsing behavior. The content mainly includes four aspects.(1) This paper introduced the Hadoop platform and data warehouse, which platform can implement high-speed and effective analysis of huge amounts of data by using the distributed storage and computation technology. And according to the characteristics of the data warehouse, this paper also put forward a research framework about the users browsing behavior.(2) Based on data warehouse, this paper built the basic data layer and theme layer, the theme layer is about the users browsing behavior.(3) By researching about the association rules algorithm and path mining algorithm, this paper put forward an algorithm of Continuous Frequent Access Path Mining Based on Hive (Hive-CFAP).(4) Based on researching about the users browsing behavior theme layer and Hive-CFAP algorithm, this paper applicated the research of the continuous frequent access path, the relationship between page views and page distances, the clustering of similar browsing user.