Design And Development Of Personalized Online Shopping System Based On User Behaiour Data

With the rapid development of Internet, the network has become a part of human life, has become an important way for people to obtain information. Due to the continuous increase of network information, people spend time to search to the related information they need from the vast amounts of information. Relying on people’s search hard to find their own goals in a short period of time. Although the search engine can help people solve this contradiction, but due to the lack of intelligent search engine and personalized recommendation, so can’t solve this problem fundamentally. Personalized user access behavior of Web based on Agent to solve this problem, the idea is proposed in this paper.This paper focuses on the browsing behavior to the user of interest to the user Webpage, for personalized user model and the current widely used rely only on the page content and the establishment method, has some enlightening. Based on summary of the research results of other scholars, Web Agent and Web data mining is studied, and according to the characteristics of user behavior, proposed a collaborative filtering algorithm based on association rule, and the Web Agent based personalized recommendation system design and prototype implementation. The work of this paper are as follows:1.This paper proposes a method to obtain a client user behavior data and data mining on behavioral data, the user access behavior and its page interested combined to build the recommendation prototype system a user access behavior of Web based on Agent personalized.2.This paper uses Web data mining of user behavior data were analyzed, because the user data server records with redundancy, and the server and the client log file differences also exist, this paper use Web data mining, user behavior data were screened and exclusion. This design recommendation prototype Web Agent personalized data processing includes three processes: Based on online monitoring, off-line learning and on-line recommendation. Online monitoring is the role of the various behavioral data, the user registration information data bound together into a valuable data source of data mining, the mining and analyzing these data, we can find association rules among them. Offline learning include content data preprocessing, data pretreatment and data preprocessing. After pre processing is finished, then choose the appropriate tools for the analysis of these models and data, select the useful rules from the data in the sea. Online recommendation module according to the user’s preference to the user recommend they might be interested in the product, and then according to the user’s reaction, system to give the unified evaluation, extracted by mining the rules of the model and comparing the current user session, personalized page generation user needs.