基于用户兴趣和项目周期的协同过滤推荐算法  

Collaborative filtering recommendation algorithm based on user interest and project cycle

作　　者：叶锡君[1] 袁培森[1] 郭小清[1] 闫智慧 何婧[1] Ye Xijun;Yuan Peisen;Gou Xiaoqing;Yan Zhihui;He Jing(School of Information Sciences and Technology,Nanjing Agricultural University,Nanjing 210095,China)

机构地区：[1]南京农业大学信息科学技术学院,江苏南京210095

出　　处：《南京理工大学学报》2018年第4期392-400,共9页Journal of Nanjing University of Science and Technology

基　　金：国家自然科学基金(61502236);国家重点研发计划重点专项(2016YFD0300607);中央高校基本科研业务费专项资金(KYZ200919)

摘　　要：协同过滤推荐算法以没有限定推荐对象类型、无需用户反馈信息等优势在众多个性化推荐算法中脱颖而出。但是现有算法缺乏对用户之间的差异和用户自身的兴趣考虑,对用户和项目之间的潜在关联考虑不充分,这些问题均会影响推荐精度。该文提出一种基于用户兴趣和项目周期的协同过滤推荐算法,该算法在计算相似度时引入用户兴趣权重UI、项目时间等因素,并采用融合因子将改进后所得用户和项目信息进行综合,获得推荐列表。对比实验得出:该算法在推荐精确度上提高了11.034%,研究结果表明:该算法可有效提高推荐精确度。Collaborative filtering recommendation algorithm takes the advantage of non-restricting recommended object type,without-user feedback information and so on,and stands out in many personalized recommendation algorithms.However,the existing algorithms don’t consider the differences between users and user interests,and the potential association between users and projects yet.These problems all affect the recommendation accuracy.This paper proposes the collaborative filtering recommendation algorithm based on user interest and project cycle,which introduces user interest weight UI,project time and other factors in the computation of similarity.This paper uses the fusion factor to integrate the improved user and project information to get the recommendation list.Finally,a comparison experiment shows that the proposed algorithm improves the recommendation accuracy by 11.034%.The research results show that the algorithm can effectively improve the accuracy of recommendation.

关 键 词：推荐算法 协同过滤 个性化 用户兴趣 项目周期 用户相似度 项目相似度 线性融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]