一种基于三部图网络的协同过滤算法

A collaborative filtering recommender algorithm based on tripartite network

推荐系统是电子商务领域最重要的技术之一,而协同过滤算法又是推荐系统用得最广泛的.提出了一种基于加权三部图网络的协同过滤算法,用户、产品及标签都被考虑到算法中,并且研究了标签结点的度对用户相似性计算的影响.实验结果表明,此算法在解决用户冷启动问题的同时,还具有较高的推荐准确性.

Recommender system is one of the most important technologies in E-commerce,and the collaborative filtering algorithm is the most widely used technique in recommender system.In this paper,we proposed a collaborative filtering algorithm based on weighted tripartite network,which takes users,items and tags into account,and we also studied the degree of tags which may affect the user-user similarity computation.The experimental results demonstrate that the algorithm can solve the cold start problem with high recommendation accuracy.