基于聚类与差异协调的协同过滤推荐算法  

Collaborative Filtering Recommendation Algorithm Basedon Clustering and Difference Coordination

作　　者：林建辉 王茜冉 詹可强[1] LIN Jian-hui;WANG Xi-ran;ZHAN Ke-qiang(College of Internet of Things and Artificial Intelligence,Fujian Polytechnic of Information Technology,Fuzhou 350000,China)

机构地区：[1]福建信息职业技术学院物联网与人工智能学院,福建福州350000

出　　处：《兰州文理学院学报（自然科学版）》2023年第6期50-54,共5页Journal of Lanzhou University of Arts and Science(Natural Sciences)

基　　金：2021年福建信息职业技术学院科研课题(Y21110)。

摘　　要：目标用户近邻的计算选择与评分预测是决定协同过滤推荐算法推荐质量与准确度的关键步骤.传统的目标近邻选择与评分预测方法未考虑用户的个人评价尺度或评价习惯,导致近邻选择不准确,影响最终的推荐结果.为了提高目标用户近邻准确度量选取的问题,本文提出基于聚类与差异协调的协同过滤推荐算法,采用结合时间权重与聚类的方式筛选出目标的近邻用户,通过评价差异因子协调评价差异完成最终预测推荐.实验结果表明,该算法能够提高推荐系统的推荐质量和准确度.The accuracy of target neighbor selection is related to the accuracy of the entire recommendation system.The traditional target neighbor selection method,because the user similarity measurement is not accurate enough,leads to the reduction of recommendation accuracy.To improve the accuracy of target neighbor measurement selection,a collaborative filtering recommendation algorithm based on clustering and difference coordination is proposed,the clustering method is used to screen the nearest neighbor users with similar goals,the difference is coordinated through the evaluation of difference factors,and the final prediction is determined by combining the prediction of the nearest neighbor users with the project popularity.The experimental results show that the algorithm can improve the recommendation accuracy and the recommendation quality of the system.

关 键 词：协同过滤 聚类 时间权重 差异协调因子 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]