基于CiteSpace的推荐系统研究可视化分析

Visualization Analysis of Recommendation System Based on CiteSpace

分析推荐系统领域研究热点主题和未来发展趋势,对提高和改进推荐结果对用户友好性具有重要意义。采用科学知识图谱分析工具CiteSpace对2001—2020年间自然科学引文数据库中推荐系统领域文献进行可视化分析和解读。研究结果显示,推荐系统领域当前热点研究有“协同过滤、个性化推荐、用户画像”,推荐系统的研究前沿发展趋势集中在:冷启动、深度学习、相似性度量。建议未来在该领域的研究树立新高度,站在多学科结合视角、加强学术间交流、增进机构间合作,消除现有技术壁垒,促进推荐系统研究积极向前发展。研究结果可为国内推荐系统研究的后续理论研究和实践操作提供参考依据。

It is a great significance for recommender system to find out the hot topics and which topics will burst in future.It is also important to improve the user experience or get better results.In order to visually analyze and interpret the search of recommendation system in WOS(Web of Science)paper data from 2001 to 2020.A scientific knowledge graph analysis tool named CiteSpace was used.The results show that the current hot topics in the field of recommender system include“Collaborative Filtering,Personalized Recommendation,User Profile”.And the current research of recommender system focus on:cold start,deep learning,similarity measurement.Hence,it is suggested that should set up a new height in the future research.Combining different disciplines.Moreover,strengthen academic exchanges,enhance inter agency cooperation,solved the existing technical barriers,and do some positive research in recommender system.It was proved that reference for the follow-up theoretical research and practical operation of domestic recommender system research.