Design And Implement Of E-commerce Recommendation System Based On Limited Resources

With the continuous development of Internet technology and E-Commerce,the way of people's life has change a lot, people can get a lot of commodity information while staying at home, however, such a large number of products make people difficult to find suitable product, which lead the information efficiency to be reduced. Facing the requirement of people want to get effective information quickly, E-Commerce recommendation system is widely used to solve the current problem. The E-Commerce recommendation system, as a guide, proposed personalized proposal which is consistent with the user preferences for users based on user's historical behavior data, to help user make better choices, so it can enhance the website satisfaction and loyalty.Current E-Commerce recommendation system have been greatly developed in theory and practice, it can recommend some personalized products according to the user's behavior. However, the current research of the E-Commerce recommendation system is lack of consideration of current frequent large-scale commercial preferential activities. Such as"double eleven",the daily commodity recommendations might be failure because of limited products. The main research points of this paper includes:(1) From the perspective of user's preferences, through the analysis of user preference of attribute, give out a user preference representation method, through the analysis of user behavior in the E-Commerce website,give out a unified description of behavior representation method. The user's preference model is based on content-based recommenders, through the analysis of user behavior, studying the calculation model of the user preference for each attribute of commodity, which gives out the content-based user preference modeling algorithm.(2) This paper puts forward the concept of the limited resources of the E-Commerce suppliers, and gives out a unified representation method for the large-scale special-sale activities. For the E-Commerce limited resources, Based on user-based collaborative filtering algorithm, improves user preference prediction formula, give out limited-resource-based collaborative filtering algorithm, correcting the recommended results with analyzed user's preferences model, which can improve the accuracy of recommendation results.Based on the research in the above aspects, this paper design and implement an E-Commerce personalized recommendation prototype system based on E-Commerce limited resources, including user preferences model and recommendation two modules, which can reasonably analyze user preference model, and recommend products based on the limited resources effectively. The research has a positive significance to solve the recommendation problem in frequent large-scale commercial preferential activities.