基于社会性网络服务的多模型推荐系统

Multi-model recommendation system based on social networking services

通过分析推荐系统中的协同过滤算法,说明单一模型的推荐系统存在着难以解决的不足,尤其是数据稀疏问题。使用基于社会性网络服务(SNS)平台的推荐系统,结合SNS平台的分享特点,提高用户参与的积极性,从而有效解决数据稀疏问题。并提出结合SNS的推荐模型、参考用户搜索行为的推荐模型、结合协同过滤的推荐模型的多模型推荐系统,以三个不同的推荐列表满足用户的不同需要。

Through analyzing collaborative filtering recommendation system, it is not hard to say that recommendation system with single model has inevitable lack, such as the sparseness problem of the data. The recommendation system based on Social Networking Services （ SNS）, containing the share feature, can improve the enthusiasm of the user, so that it can solve the sparseness problem of the data in certain degree. This paper proposed a kind of recommendation system with three different models. They are the model combined SNS, the model referred users＇ searching trace and the model combined collaborative filtering. Recommendation system with above models would supply three recommendation lists to fit different users＇ different need.