Design And Implementation Of Web Recommendation System

To face a lot of information and diversification of information, more and more people can get information. But it is quite difficult to get the information quickly and efficiently and accurately. The search engine appears to solve the user’s retrieval needs, but not in a timely manner for the user initiative to provide the information needed. When the user uncertain Keywords describe the information required, and not be able to search; or users themselves do not want to search, just want to get some information recommended. Then it is needed a more active than the search engine system to meet these needs, the intelligent recommendation service system was born. Therefore, recommendation technology is also increasingly becoming an important research topic.In this paper, a Web recommendation system was designed and implemented for the user demand for information recommendation. Key technology of the system is the collaborative filtering algorithm based on singular value decomposition. And the system preprocesses information of network with the use of feature selection methods in text categorization, which also for processing the collected pages and pages of potential users interested. While this paper introduced poly class analysis algorithms to solve the problem of server load and the results of recommending not accurate. The final through users’ feedback and users’ using time periods optimize the recommended results list to improve the recommendation result accuracy. This makes recommendation results more in line with the needs of users and more humane. This design of Web recommendation system is user-oriented, combining user interest and hybrid recommender system using collaborative filtering recommendation algorithm based on singular value decomposition. It is considered for the recommended system that the page the user is viewing, user interests and the records of user history access, which combined with user using time to provide users with Web referral service. The ultimate realization of the recommendation system can find the user’s interests and hobbies, then recommend information consistent with users’ hobbies or items to the target users, with reference to user feedback and user usage periods to provide smart recommended.Firstly the paper reviewed that Web recommendation system research background, status and content as well as in practice. The paper presents the overall framework and outline design of a Web recommendation system based on Web recommendation system requirements analysis, the detailed design and coding to achieve module and a detailed introduction to key technologies and methods and instructions. The last experimental testing of web recommendation system, and pointed out that the web recommendation system needs further improvement and refinement aspects.