摘要
传统的位置服务推荐方法主要存在数据稀疏、推荐精度低、无法满足用户个性化需求等问题,通过考虑用户的位置信息和语义信息,提出一种基于语义分析的位置服务推荐方法。首先,设计一种位置⁃语义转换算法,提取数据集中位置数据对应的语义信息,针对用户的移动轨迹序列生成语义轨迹序列;其次,从用户的语义轨迹序列中提取子序列,通过子序列匹配算法,挖掘共同子序列,发现与目标用户具有潜在好友关系的相似用户;最后,中心服务器将相似用户所访问的实际位置信息推荐给目标用户,完成位置服务推荐。实验证明所提出的基于语义分析的位置服务推荐方法不仅能够有效发现相似好友,而且能够提高推荐精度,更重要的是,该方法能够将从未访问过的位置推荐给目标用户,满足用户的个性化需求。
The traditional location service recommendation methods have some deficiencies,such as sparse data,low recommendation accuracy and unable to meet the personalized needs of users,so a location service recommendation method based on semantic analysis is proposed by taking into account the location information and semantic information of users.A position to semantic(PTS)conversion algorithm is designed to extract the semantic information corresponding to the location data in the data set,so as to generate the semantic trajectory sequence for the user′s moving trajectory sequence.The sub⁃sequences are extracted from the semantic trajectory sequence of the user,and then the common sub⁃sequences are mined by the sub⁃sequence matching algorithm,so as to find the similar users who have potential friend relationship with the object users.The central server recommends the actual location information accessed by similar users to the object users to complete location service recommendation.The experiment results show that the proposed location service recommendation method based on semantic analysis can not only find similar friends effectively,but also improve the recommendation accuracy.More importantly,this method can recommend the locations that the object users have never visited to them and meet their personalized needs.
作者
余丽萍
朱亮
刘啸威
YU Liping;ZHU Liang;LIU Xiaowei(Zhengzhou University of Light Industry,Zhengzhou 450001,China)
出处
《现代电子技术》
2022年第9期98-104,共7页
Modern Electronics Technique
基金
国家自然科学基金资助项目(61902361)
河南省科技攻关项目(212102210095)。
关键词
位置服务推荐
语义分析
位置⁃语义转换
数据提取
位置信息推荐
数据处理
location service recommendation
semantic analysis
PTS conversion
data extraction
location information recommendation
data processing
作者简介
余丽萍(1984-),女,河南郑州人,工学硕士,主要研究方向为服务计算;朱亮(1987-),男,河南郑州人,工学博士,讲师,主要研究方向为服务推荐与隐私保护;刘啸威(1998-),男,河南郑州人,在读硕士生,主要研究方向为服务推荐与隐私保护。
