摘要
相似度计算是协同过滤推荐算法的基础,但由于用户之间共同评价项目数量稀少,导致传统的协同过滤算法无法准确计算用户之间的相似度,从而造成推荐质量不佳。通过在Pearson相似度计算公式中加入用户之间联系的惩罚因子,并在此基础上与评分信息熵进行融合,提出一种新的用户之间相似度计算方法。实验结果表明,该算法能够更准确地计算用户之间的相似度,从而提高推荐结果质量。
The similarity calculation is the basis of the collaborative filtering recommendation algorithm.However,due to the scarcity of common evaluation items among users,traditional collaborative filtering algorithms cannot accurately calculate the similarity between users,resulting in poor recommendation quality.By adding the penalty factor of the connection between users into the Pearson similarity calculation formula,and fusing with the rating information entropy on this basis,a new calculation method for the similarity between users is proposed.Experimental results show that the algorithm can calculate the similarity between users more accurately,thus improving the quality of recommendation results.
作者
张洁
李港
ZHANG Jie;LI Gang(School of Computer Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China)
出处
《南京邮电大学学报(自然科学版)》
北大核心
2021年第2期71-76,共6页
Journal of Nanjing University of Posts and Telecommunications:Natural Science Edition
基金
国家重点研发计划(2018YFB1500902)
南京邮电大学校级科研基金(NY219122)资助项目。
关键词
协同过滤
相似度计算
评分信息熵
评分预测
collaborative filtering
similarity calculation
rating information entropy
rating prediction