Researcn And Implementation Of Wine Busines Platform Based On Collaborative Filtering Recommendation

With the rapid development of electronic commerce, large scale comprehensive shopping website has been growing, all kinds of niche shopping sites also vigorous development. Wine electricity sector attention by VC constantly, the competition is becoming increasingly fierce, the problem is more prominent; As the increasing number of users and goods continues to increase, the user found in the thousands of goods be fond of your product are more and more time-consuming. Therefore, this paper constructed a wine business platform to meet the current demand, through the discovery of user interest to personalized recommendation for users to enhance user satisfaction and website competitiveness. In the personalized recommendation technology, collaborative filtering recommendation algorithm is widely used in E-commerce field, but the algorithm has many problems, so the algorithm is studied in this paper.In this paper, a detailed analysis based on users and item collaborative filtering algorithm and the advantage and disadvantage of the two methods and the existing problems compared, combined with the characteristics of the wine itself, has chosen the user-based collaborative filtering algorithms as a basis for improvement. Fusion is proposed a collaborative filtering recommendation algorithm based on item properties, the traditional collaborative filtering algorithms in a user item rating and item attribute combination, without any increase in user feedback rating, the evaluation of the project attribute value is converted to the fine granularity by the users of the project coarse granularity evaluation, the latter is relative to the former to a certain extent reduces the score matrix sparse degree, and the structure is more stable. On this basis, the users on different attributes similarity calculation, get comprehensive weighted similarity and predict ratings. In this paper, considering the cold start problem caused by the addition of the new project, the combination of the weighted user similarity prediction score and the project attribute similarity prediction score is recommended for users. Finally, the crawler program to obtain a wine sales site wine attributes, user evaluation data, finishing the experimental data set. The validity of the proposed algorithm is verified by the experimental results.Designed a wine sales oriented business website, set up to meet the current needs of the business website technology architecture, to achieve a complete front desk services and background management functions. According to the website demand orientation with regional characteristics, realize the city card and access to the public payment function, improve the loyalty of the users of the site.