摘要
兴趣点推荐是基于位置社交网络中的研究热点之一。首先对从Web of Science收集的兴趣点推荐研究文献进行了分析;然后分析了影响兴趣点推荐的多种因素,并在分析传统兴趣点推荐方法基础上重点从用户历史签到信息建模和用户社交生成信息提取两个方面对基于深度学习的兴趣点推荐方法进行了分析;最后,对未来可能提高兴趣点推荐效果的研究方向进行了展望。
Point of interest recommendation is one of the research hotspots in location-based social networks.Firstly,this paper analyzed the references about point-of-interest recommendation research from Web of science,then analyzed the factors that influenced point-of-interest recommendation.Furthermore,based on the analysis of traditional methods of point-of-interest recommendation,this paper mainly analyzed the methods of point-of-interest recommendation based on deep learning from two aspects:modeling of user historical check-in information and extraction of user social generated information.Finally,it looked forward to future research directions that might improve the performance of point-of-interest recommendation.
作者
李征
黄雪原
袁科
Li Zheng;Huang Xueyuan;Yuan Ke(College of Computer&Information Engineering,Henan University,Kaifeng Henan 475004,China;Henan Key Laboratory of Big Data Analysis&Processing,Henan University,Kaifeng Henan 475004,China;Henan Spatial Information Processing Engineering Laboratory,Henan University,Kaifeng Henan 475004,China)
出处
《计算机应用研究》
CSCD
北大核心
2022年第11期3211-3219,共9页
Application Research of Computers
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资助项目(182102410063)
河南省高等学校重点科研项目计划(23A520016)。
关键词
兴趣点推荐
数据稀疏
访问序列模式
注意力机制
图嵌入
point-of-interest recommendation
data sparsity
access sequential pattern
attention mechanism
graph embedding
