基于最近邻居优化选取方法的协同过滤推荐算法

Collaborative Filtering Recommendation Algorithm Based on Nearest Neighbor Optimal Selection Method

针对传统的协同过滤推荐算法推荐精度低和数据稀疏的问题,提出基于最近邻居优化选取方法的协同过滤推荐算法.首先,提出一种用户可用度计算模型,根据其他用户对目标用户的可用度计算结果,选取最近邻居候选集.然后,提出一种用户信任度计算模型,计算目标用户对最近邻居候选集中用户的信任度,进而选取目标用户的最近邻居.最后,根据最近邻居的评分情况,得到目标用户的推荐.实验结果表明,该算法提高了推荐精度,而且有效地改善了不同稀疏程度数据上的推荐效果.

To solve the problems of the low recommendation precision and data sparseness in traditional collaborative filtering, a collaborative filtering recommendation algorithm based on nearest neighbor optimal selection method was proposed. Firstly, the availability computing model was designed to calculate availability between users and on the basis of the availability, the alternative nearest neighbor set of target user was selected. Then the trust degree computing model was designed to calculate trust degree according to the ratings of alternative nearest neighbors, and the nearest neighbor set of target user was chosen based on the trust degree between users. Finally, the target user＇s recommendation was obtained according to the nearest neighbors＇ ratings. Experimental results show that the proposed algorithm not only can improve the recommendation precision, but also can efficiently improve the recommendation quality on different sparsity data.