融合PCA降维和均值漂移聚类的协同过滤推荐算法

A collaborative filtering recommendation algorithm combining PCA dimension reduction and mean shift clustering

为了解决传统的协同过滤推荐算法中评分矩阵稀疏及近邻搜索耗时长导致的推荐准确性及时间效率有待提升的问题,设计了一种融合PCA降维和均值漂移聚类的协同过滤推荐算法PMCF。该算法用主成分分析法PCA保留最能代表用户兴趣的维度,以缓解评分矩阵稀疏问题;用均值漂移聚类算法在降维后的低维向量空间上对用户聚类,以减小目标用户最近邻的搜索范围。在Movielens数据集和HetRec2011⁃Movielens⁃2k数据集上的实验结果表明,PMCF算法能够有效地提升推荐结果的准确性,同时具有较高的时间效率。

Traditional collaborative filtering recommendation algorithms have faced the problems of lower accuracy and time efficiency,due to the sparse score matrix and long time⁃consuming neighbor search.This paper proposes a collaborative filtering recommendation algorithm PMCF combining principal component analysis(PCA)dimension reduction and mean shift clustering.In this algorithm,PCA is used to retain the dimension that can best represent the user’s interest,so as to alleviate the problem of sparse score matrix.The mean shift clustering algorithm is used to cluster users in the reduced dimension vector space to reduce the search range of the nearest neighbor of the target user.The experimental results on Movielens dataset and HetRec2011⁃Movielens⁃2k dataset show that the PMCF algorithm can effectively improve the accuracy of recommendation results,and has high time efficiency.