一种优化的组合协同过滤算法

An Improved Combining Collaboration Filtering Algorithm

改进了传统协同过滤算法中最近邻搜索这一关键步骤,提出了一种结合概念层次和用户局部兴趣相似的协同过滤算法,使推荐系统在用户矩阵整体稀疏局部密集时依然能产生较好的推荐.该算法应用于基于iPhone平台开发的EatMe菜肴推荐系统,实验证明改进算法比传统协同过滤算法有更高的查全率.

This paper improved on computing neighbor users for active users,which was the crucial step of traditional collaboration filtering algorithm.It proposed a collaboration filtering algorithm combining concept hierarchy and users′ partial similarity so that it could still work on the sparse but partial dense user matrix.This algorithm was applied in EatMe dish recommendation system based on iPhone,and the experimental results showed that the improved algorithm can provide better recall than traditional collaboration filtering algorithm.