融合用户相似度的协同过滤算法研究  

Research on Collaborative Filtering Algorithm Based on User Similarity

作　　者：贾雯 李香[1] Jia Wen;Li Xiang(College of Computer Science and Technology,Harbin Engineering University,Harbin Heilongjiang 150001,China)

机构地区：[1]哈尔滨工程大学计算机科学与技术学院,黑龙江哈尔滨150001

出　　处：《信息与电脑》2020年第14期74-76,共3页Information & Computer

基　　金：国防基础科研计划(项目编号:JCKY2017603B006);面向国家重大需求培育计划中央高校基金重大项目“高负荷跨声速涡轮高效通流构型及多维耦合快速设计方法研究”(项目编号:HEUCFP201720)。

摘　　要：降低相似度计算偏差是提升系统推荐准确性的核心。针对数据稀疏、项目热门程度等影响相似度计算准确性的因素,笔者在Pearson相关系数基础上,提出一种融合Jaccard系数和频率加权Pearson系数的算法。改进后的算法考虑了用户间共同评分项数量占比,且对热门项目添加相关惩罚权重因子,降低其对相似度计算产生的影响,进而提升相似度计算的准确性。实验结果表明,所提算法平均绝对误差低于传统算法,有效提高了系统推荐的准确性。Reducing the similarity calculation deviation is the core of improving the system recommendation accuracy.Aiming at the factors that affect the accuracy of similarity calculation,such as data sparseness and item popularity,an algorithm combining the Jaccard method and frequency weighted Pearson coefficient is proposed based on the Pearson correlation coefficient.The improved algorithm takes into account the proportion of the number of common scoringitems among users and adds relevant penalty weight factors to popular items to reduce their impact on the similarity calculation,thereby improving the accuracy of similarity calculation.Experimental results show that the average absolute error of the proposed algorithm is lower than that of the traditional algorithm,which effectively improves the accuracy of system recommendation.

关 键 词：推荐系统 协同过滤 相似度 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]