面向C2M个性化定制的协同过滤推荐算法研究

Research on Collaborative Filtering Recommendation Algorithm for C2M Personalized Customization

摘要: 协同过滤算法是现今电商模式下最有效和最广泛的推荐方法，但是传统算法仅考虑用户兴趣，忽略项目信息的影响。特别在C2M (用户直连制造)模式下，借助于工业定制平台，项目的有效信息量倍增，项目信息对于用户的个性化定制有重大的影响。针对这个问题，本文构建用户满意度模型来预测项目的信息对用户的影响，结合传统协同过滤算法，综合考虑用户兴趣和满意度来产生推荐结果。通过实验结果表明，改进协同过滤算法的具有可行性，并提高了推荐的有效性。

Abstract: Collaborative filtering algorithm is widely used in e-commerce mode. Traditional algorithms only take into account the interest of users, which ignore the impact of item information. Especially in the C2M (Customer-to-Manufacturer) mode, with the help of the industrial customization platform, the item information is increased rapidly and makes a significant impact on the user’s personalized customization. To solve the problem, this text builds a customer satisfaction model to predict the impact of item information on users, and combines traditional collaborative filtering algorithm to generate recommendation results that take into consideration the interest and satisfaction of users. Experimental results show that the improved collaborative filtering algorithm is feasible and im-proves the effectiveness of the recommendations.