摘要
用户在无线网络间漫游时产生了大量的行为数据。这些数据蕴含着用户的生活轨迹,轨迹越相似的用户具备亲密社会关系的可能性越大。传统方法通过比较两条语义轨迹中的最长公共子序列来挖掘用户之间轨迹的相似程度。但这种算法忽视了轨迹的时序性和轨迹点的连续性。为此,提出了一种基于时间特征和空间特征的轨迹相似度计算方法,从时空两个维度计算用户的轨迹距离,并依据轨迹相似度对用户聚类,挖掘不同时间切片下的聚类结果,对亲密度更高的用户对进行“共同漫游行为”的画像。实验结果表明,在无线漫游场景下,该方法可以较为准确地衡量用户之间的相似度,在找出具备社会关系的用户方面具有较好的效果,并能可视化用户间的共同漫游行为。
Users generate a lot of behavior data when roaming under wireless network.These data contain users’trajectories.The more similar the trajectory is,the more likely users have intimate social relationship.Traditionally,the similarity of trajectory between users is mined by comparing the longest common subsequences of two semantic trajectories.However,these kinds of algorithms do not focus on sequential properties and continuity of trajectories.In this paper,a method of trajectory similarity computation based on spatio-temporal character is proposed.The trajectory similarity of users is calculated from two dimensions of time and space.Users are clustered based on the trajectory similarity.Then clustering results are mined under different time slices,drawing a persona for users who have higher intimacy as a“common roaming behavior”persona.Experimental results show that this method can accurately measure the similarity between users,performs better to find social relationship,and can visualize the common roaming behavior among users in wireless roaming scenarios.
作者
常祎祎
王劲松
Chang Yiyi;Wang Jinsong(School of Computer Science and Engineering,Tianjin University of Technology,Tianjin 300384;Tianjin Key Laboratory of Intelligence Computing and Novel Software Technology,Tianjin 300384;National Engineering Laboratory for Computer Virus Prevention and Control Technology,Tianjin 300457)
出处
《高技术通讯》
EI
CAS
北大核心
2019年第9期862-868,共7页
Chinese High Technology Letters
基金
国家重点研发计划(2018YFC0831405)
天津市自然科学基金重点(18JCZDJC30700)资助项目
关键词
校园网
无线漫游
时空数据
轨迹相似度
聚类
campus network
wireless roaming
trajectory similarity
spatio-temporal data
clustering
