SVDSLIM推荐算法研究

Research on SVDSLIM Recommendation Algorithm

Pure SVD推荐算法通过重构分解矩阵而获得推荐结果,稀疏线性(SLIM)算法侧重于通过挖掘物品之间的线性关系而产生推荐效果。为了提高Top-N推荐质量,文章提出了一种基于Pure SVD、SLIM的混合推荐算法—SVDSLIM算法。该算法会对Pure SVD、SLIM算法的推荐结果赋予不同的权重,然后将处理过的推荐结果作为新的推荐列表。实验结果表明,SVDSLIM算法在各种数据集上推荐质量要优于其他算法。

Pure SVD recommendation algorithm obtains the recommendation result by reconstructing the decomposition matrix,and the sparse linear（SLIM） algorithm focuses on the recommendation of mining the linear relationship between items. In order to improve the quality of Top-N recommendation, this paper proposes a hybrid recommendation algorithm-SVDSLIM based on Pure SVD and SLIM. The algorithm will give different weights to the recommended results of the Pure SVD and SLIM algorithms, and then use the recommended results as a new recommended list. The experimental results show that the SVDSLIM algorithm is superior to other algorithms in various data sets.