The Design And Implementation Of Personalized Recommendation System Based On Web Log

With the popularization of computer network, more and more people spend their time on the Internet. The development and promotion of HTML5and other technology will promote the development of web applications. Web data mining and personalized recommendation have become a hot research. As we can see, in the field of electronic commerce, this technology has been widely applied. It stimulates the buyers’ desire to buy the goods through recommending the goods. There are a lot of web resources which can be recommend not only commodity.This paper puts forward a web page personalized recommendation system on the basis of the utilization of the web log data mining.At first, this paper introduces a scheme of distributed log data collection. This scheme can be applied to the collection and storage of massive distributed data.The collected data will be used as the data source of Web log mining. Then, the paper introduces the pretreatment process of Web Log Mining. It provides detail design of every step of the pretreatment process. Extracting the users’ accessing information from the log lay a good foundation for the data mining. Then this paper analysis the commonly used clustering algorithm. Because the web user clustering and webpage clustering have the obvious feature of fuzzy, this paper will use fuzzy clustering technology in the web log clustering process, as the foundation of personalized webpage recommendation. The off-line clustering analysis is collecting the web data of web users and generating clustering results including user clustering and item clustering using the the clustering algorithm above. In the on-line real-time recommendation part, the system determines the users’clustering or find the item clustering. Then the system recommend the user of the items their maybe interested in and displaying them on the webpage their visit, so as to achieve the purpose of personalized recommendation.Finally, the recommendation system based on the design method above is implemented. The system is tested using the actual log data collected. A background system is also implemented to observe the experimental results, using specific evaluation methods. The feasibility of every process is validated in the system.