摘要
为了提高个性化推荐的准确度,提出一种基于全局相似度的在线资源个性化推荐算法.首先分别基于用户和在线资源对其各自相似性进行计算,运用K-means聚类方法根据对用户偏好和在线资源属性及评分的相似性计算结果对其进行划簇,实现用户与在线资源聚类划分,以此实现精准的个性化在线资源推送.实验结果显示,本文方法推荐的最小平均绝对误差为0.77,查准率随着数据覆盖率的增加可达到60%以上,推荐耗时基本稳定在20 s以内.在推荐准确度、查全率以及效率方面均有良好表现.
In order to improve the accuracy of personalized recommendation,this paper proposes a personalized recommendation algorithm for online resources based on global similarity.Firstly,the similarities between users and online resources are calculated respectively,and K-means clustering method is used to classify the users and online resources based on the similarity calculation results of users’preferences,and the attributes and scores of online resources,so as to offer the accurate personalized online resources.The experimental results show that the recommended minimum mean absolute error of the proposed method is 0.77,that the precision rate can reach more than 60% with the increase of data coverage rate,and that the recommendation time is basically stable within 20 s.Therefore,the method has a good performance in recommendation accuracy,recall rate and efficiency,which has a certain research value.
作者
郑辉
ZHENG Hui(School of Computer and Art,Anhui Technical College of Industry and Economy,Hefei 230051,China)
出处
《常熟理工学院学报》
2021年第5期75-80,共6页
Journal of Changshu Institute of Technology
基金
安徽省高等学校自然科学研究重点项目“基于HBase的动态可配置数据采集整合与提取复用系统的应用研究”(KJ2018A0764)。
