EC Intellectualized Recommendation System Research

With the development of internet and electronic commerce ( EC ), the structure of EC system is becoming more and more complex. And the users feel confused facing so many items to choice, because it is hard for them to find out what they really need and where what they need are. The EC recommendation systems can interact with the users, recommend items to the users and help the users to find out where the items they really want. So under the recommendation systems help, the buying procedure is very smooth. Facing the tough competition, the EC recommendation system can help enterprises to hold the customers efficiently, boost the sale volume and enhance the competition power.The recommendation system has a amazing development prospect. It is becoming a hot issue among the IT technology, and it attracts more and more attention of observers.EC recommendation system has made a great advance both in theory and in practice. But it also faces a series of challenges when its scale becomes bigger. Aim at these challenges, the paper studies EC recommendation system from three aspects:Firstly, the recommendation system the paper studies adapts the mechanisms from voting theory. Most recommendation system adapts social filtering approach mechanisms, but the social filtering approach mechanisms can not resolve the problem that user's preferences are often conflicting well. In the other hand, the mechanisms from voting theory can. Furthermore, the excellent recommendation system should rank the items according to user's preferences and provide the result to user. The mechanisms from voting theory also do it well.Secondly the paper adapts naive Bayes classifier to analyze the contents of the contents of items' summaries. Base on the naive Bayes classifier, the recommendation system can track the user's preferences change, then it can change the user modeling accordingly. This can reduce the use's workload of setting preferences information and lighten the burden of user.Finally, the paper analyzes the structure, the function and the algorithms of watch recommendation system.