Design And Implementation Of UGC Intelligent Recommendation System Based On Social Intercourse Electronic Commerce

User Generated Content (UGC) is the main information mode of social intercourse electronic commerce. UGC presents fragmentation, massive, disorderly characteristics, and how to use the UGC data from the social electricity suppliers to find the commodity has become an urgent need to solve the problem. Currently, many recommendation systems are based on the specified electricity supplier platform. Based on the recommendation system from the social business platform of UGC data, the data is messy, the subjective strong, if you use the proposed algorithm to recommend the current, it is obviously not suitable for use. This paper presents a new intelligent recommendation algorithm based on social electricity suppliers UGC data. In order to improve the accuracy of the recommendation system, this study proposes a new algorithm based on multi object fuzzy classification algorithm, which is based on the initial recommendation of the potential factor algorithm, and improves the traditional collaborative filtering algorithm based on user’s membership.Compared with the traditional electronic business platform, UGC electric business platform only have a small part of the total score, and its mathematical abstraction is a sparse matrix. The traditional Slope-One algorithm is not suitable for UGC recommendation system. For a UGC business platform, according to the characteristics of similar preferences of the user group can better guess the user’s preferences of unfamiliar goods, this study combines the characteristics of UGC platform, and focus on solving intelligent recommendation based on user clustering.Through the analysis of the algorithm and system function, this paper analyzes the system architecture design and functional analysis of UGC intelligent recommendation system based on social electricity suppliers. The paper studies and solves two new problems of the new user and the product. On the other hand, it also makes the evaluation of the new product. Finally, the experimental data is realized by the Mahout Apache of the open source project and the real movie score data obtained from the Group Lens project, and the validity of the algorithm is verified.