一种缓解互惠推荐系统中数据稀疏性的算法  

An Algorithm for Relieving Date Sparsity in Reciprocal Recommendation System

作　　者：殷方勇[1,2] 王红[1,2] 王吉华[1,2] 

机构地区：[1]山东师范大学信息科学与工程学院,山东济南250014 [2]山东师范大学山东省分布式计算机软件新技术重点实验室,山东济南250014

出　　处：《济南大学学报（自然科学版）》2017年第1期48-54,共7页Journal of University of Jinan(Science and Technology)

基　　金：国家自然科学基金项目(61472233)

摘　　要：为解决以在线交友为代表的互惠推荐系统中数据稀疏性问题,根据LMa Fit算法提出一种改进评分矩阵的互惠推荐算法,该算法改进了传统评分矩阵填充的单向性与融合相似度计算。结果表明,与基于项目的协同过滤推荐算法和基于内容和协同过滤的混合算法相比,改进评分矩阵的互惠推荐算法在准确率、召回率和调和平均数值方面有明显改进,该算法不仅改善了数据稀疏性的问题,而且推荐质量也明显优于其他算法。In order to solve the problem of data sparsity in the reciprocal recommendation system, which was represented by online dating, based on LMaFit algorithm, a reciprocity recommendation algorithm based on improved scoring matrix was proposed, which improved the one-way filling of the traditional scoring matrix and fusion similarity calculation. Compared with the collaborative filtering recommendation algorithm based on item and hybird algorithm based on content and collaborative filering, the reciprocity recommendation algorithm based on improved scoring matrix has obvious improvement in precision,recall and mean value. This algorithm improves the data sparsity problem and the recommended quality is also significantly outperforms other algorithms.

关 键 词：数据稀疏性 互惠推荐 矩阵填充 融合相似度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]