基于隐式反馈的推荐研究综述  

Survey of Recommendation Study Based on Implicit Feedback

作　　者：李征[1,2,3] 金迪 黄雪原 袁科[1] LI Zheng;JIN Di;HUANG Xueyuan;YUAN Ke(Henan University School of Computer&Information Engineering,Henan Kaifeng 475004,China;Henan University Henan Engineering Laboratory of Spatial Information Processing,Henan Kaifeng 475004,China;Henan University Henan Key Laboratory of Big Data Analysis and Processing,Henan Kaifeng 475004,China)

机构地区：[1]河南大学计算机与信息工程学院,河南开封475004 [2]河南大学河南省空间信息处理工程实验室,河南开封475004 [3]河南大学河南省大数据分析与处理重点实验室,河南开封475004

出　　处：《河南大学学报（自然科学版）》2022年第3期305-319,共15页Journal of Henan University:Natural Science

基　　金：国家自然科学基金资助项目(61402150,61806074);河南省科技攻关计划项目(182102410063)。

摘　　要：隐式反馈在推荐系统中可以有效地表达用户-项目之间丰富的交互信息.根据隐式反馈数据量大且易获取的特点,与推荐算法进行融合,不仅可以有效缓解用户偏好表达不完整的问题,还可以提高推荐准确度.因此,研究人员对隐式反馈推荐开展了大量研究,但缺少相关研究的综述分析.本文根据收集到的文献对隐式反馈推荐热点词汇进行了分析,按照研究问题-研究内容的思路重点对隐式反馈推荐研究进行了综述,并阐述了隐式反馈推荐的应用以及未来可能的研究方向.Implicit feedback can effectively express the rich information of user-item interaction in a recommendation system.Based on the characteristics of large amount of implicit feedback data and easy access,integration with recommendation algorithms can not only effectively alleviate the problem of incomplete expression of user preferences,but also improve recommendation accuracy.Therefore,researchers have conducted a large number of studies on implicit feedback recommendation,but there is a lack of review analysis of these studies.Firstly,the author analyzed the hot terms of implicit feedback recommendation based on the literature collected from Web of Science;then focused on the review of implicit feedback recommendation research according to the research problem-research content;finally,the author described the applications of implicit feedback recommendation and possible future research directions.

关 键 词：隐式反馈 用户偏好 单类协同过滤 辅助信息 模型融合 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]