Consumer Psychology Based Recommended Algorithm In E-Commerce System

In the area of E-commerce, with the increasing commodity information, the users seem even more difficult to find the goods they need. If the e-commerce sites require a higher efficiency, it must be to better meet the needs of users. Personalized recommendation system can accomplish this task well. To make it easier for users to find the goods they need. Personalized recommendation system can enhance customer loyalty and create more value for enterprise.This thesis study deeply from e-commerce data warehouse, personalized recommendation system and consumer psychology, and then analyzes the shortcomings of the current personalized recommendation algorithm; consumer psychology is used in personalized recommendation systems.First, for the problem of consumer's attitude of goods can not be described accurately by feature- based recommendation algorithm existing. This thesis presents a personalized recommendation algorithm based on multi-attribute model of attitude of user salient belief, user salient belief and multi-attribute model of attitude-based data weight are proposed. The algorithm describes user's psychology from different view, for this reason, the recommendation result is more satisfied user's needs, reflects the different characteristics of consumer goods different consciousness.Second, by further study found that multi-attribute model of attitude of user salient belief recommendation algorithm has not take into account the user understanding of changes in commodity features. For this reason, thesis also present a recommendation algorithm to adapt to changes in user salient belief, user recent interest weight and user history interest weight are proposed. This algorithm improves the result more exactly. To effectively solve the user's salient beliefs changes.The last, to verify the proposed personalized recommendation algorithm feasibility and correctness, in this thesis, experimental model established, and made with the Java implementation of the recommendation algorithm, achieve the desired results.