基于PCA降维协同过滤算法的改进

随着电子商务网站用户与商品数目的增加,使用户-项目评分矩阵成为高维稀疏矩阵,使协同过滤算法的质量降低.为此,采用主成分分析法对用户-项目评分矩阵进行降维处理,改善输入数据的稀疏性.实验结果表明,与几种典型的协同过滤算法比较,改进后的算法推荐质量有明显提高.

Improvement on Collaborative Filtering Algorithm Based on PCA Default-Values

With the rapid lincrease of users and commodities,user-item rating matrix has become the High-dimensional sparse matrix,causing collaborative filtering algorithm being low quality.Using the principal components analytic method to reduce the dimension of the user-item rating matrix so as to improve its sparsity.The experimental results demonstrated that compared with other collaborative filtering algorithm,recommendation quality of this algorithm is improved obviously.