改进预测评分矩阵的协同过滤算法

Collaborative Filtering Algorithm for Improving Predictive Scoring Matrix

为解决传统协同过滤推荐算法数据稀疏、可扩展性差等问题,采用改进预测评分矩阵的协同过滤算法。首先使用基于线性回归分析的加权Slope One算法,在传统Slope One算法中加入可信度,提高共同评分基数;然后采用网上标准数据集movielens作为测试数据,结合协同过滤算法进行top-N推荐。实验结果表明,使用改进预测评分矩阵的协同过滤算法的MEA较小,在近邻数大于25时达到0.74,表明该算法改善了传统协同过滤算法数据稀疏、扩展性差问题,降低了推荐误差,提高了推荐系统准确度。

In order to solve the problems of sparse data and poor scalability in the traditional collaborative filtering recommendation algorithm,this paper adopts a collaborative filtering algorithm that improves the predictive scoring matrix.Firstly,by using the weighted slope one algorithm based on linear regression analysis,the credibility is added to the traditional Slope One algorithm to improve the common scoring base.Then,the online standard data set movielens is used as the test data combined with the collaborative filtering algorithm for top-N recommendation.The experimental results show that the MEA using the collaborative filtering algorithm with improved predictive scoring matrix is smaller,reaching 0.74 when the neighbor number is greater than 25.The results show that the proposed algorithm improves the sparseness and poor scalability of traditional collaborative filtering algorithms,reduces the recommendation error,and improves the accuracy of the recommendation system.