The Researcher Of GEP-based Personalized Recommendation Technology In Web Site

If the past of the Internet is a popular search technology era, the future of the Internet will belong to personalized recommendation technology. Web personalized recommendation technology focus is on each workstation recommendation algorithm using different stations intelligent Web recommendation.A complete recommend system consists of three parts: behavior recording module, model analysis module and recommendation module.Behavior recording module for recording acts can reflect the behavior of the user preferences, Web personalized recommendation system tracks user behavior depends mainly on web servers WebLog (log records). With the ever-changing internet technology, cloud computing era has arrived, compared with Web Internet programming and Web data mining is concerned, Web server logs user activity log recording technology seems backward, but also in the record time of static sites. This paper studies the system, and proposed a new based Server Session log format. This log format can not only meet the server- traditional Web technologies for Web Services technology website under the user access to the records required. The new log format can be more accurate, appropriate and efficient records of user access to Web sites. Furthermore, experiments show that the log is more suitable for the needs of Web usage mining can be effective and more efficient access to the excavation site to the user model for Web site construction and development. Model analysis module is implemented on user behavior analysis of records with different recommendation techniques (algorithms) to establish model describes the user's preference information. Personalized Recommendation There are three types: rule-based filtering, content-based filtering technology, based on collaborative filtering technology.Current rules filtering technology mainly through association rule mining to discover information through the user's static characteristics and dynamic properties to recommend information. And then, Web server logs in order to tap the hot set and its associated Web page to access rules, provide referral services for primary, a generalization of the GEP-based mining multilevel association rules (mining Multiple-layers Association rule using Generalizing- based GEP, MAGGEP) algorithm.Collaborative filtering technology is by far the most successful personalized recommendation technology. At present mainly for collaborative filtering problems, conduct research in different ways. To further address the expansion of collaborative filtering technology performance problems,currently, more effective way to present the user ratings data in a cluster analysis to do. this article based on density-based clustering, recommend the gene expression programming technology to personalized recommendation , proposed a clustering based on density of GEP (Density-based methods GEP-Cluste, DGEPC) algorithm to solve the nearest neighbor problem, clustering users for services of personalized recommendation system.