Research On Collaborative Filtering Technologies In Agent Recommender System In E-commerce

Whit the development of e-commerce and the sharp increase in the amount of information, recommender system has moved from research to applications and the commercial value of Recommender System in E-Commerce has become obvious gradually. At the same time, the research of recommendation technology has been paid gradually.In this thesis, through the introduction of Recommender System in E-Commerce and the relevant theory and technology of it, it analyzes the present problems in Recommender System in E-Commerce, proposes Agent Recommender System in E-Commerce (ARS) model. At the same time the application background, application fields as well as flexible of the model are analyzed to prove the feasibility of this system,and introduces its structure, function, running processes.In this thesis, on the basis of an analysis of defects in the traditional collaborative filtering recommendation and combining the characteristics of ARS model, this thesis improves on the traditional collaborative filtering recommendation, and proposes the Dominant Rating-based and Invisible Access-based collaborative filtering recommendation, describes the process and implementation of the algorithm, and proves its validity.Finally, this thesis summarizes the work studied and proposes the direction of further research.