| 应用时间滑动窗口模型的轨迹相似性研究 |
| |
| 引用本文： | 韩奕,杜彦辉,陈庆港,芦天亮.应用时间滑动窗口模型的轨迹相似性研究[J].北京理工大学学报,2021,41(11):1207-1214. |
| |
| 作者姓名： | 韩奕 杜彦辉 陈庆港 芦天亮 |
| |
| 作者单位： | 中国人民公安大学信息技术与网络安全学院,北京100038 |
| |
| 基金项目： | 国家重点研发计划"网络空间安全"重点专项（2017YFB0802804）；国家自然科学基金资助项目（61602489） |
| |
| 摘 要： | 针对社交网络用户个人信息难以获取、公开信息不完整、不通用甚至内容虚假的问题，选择了普适性强，且能客观、真实反映用户行为习惯的位置数据作为相似性分析依据，对新浪微博、滴滴打车进行位置数据采集，形成两个高价值且具有国内网民特色的数据集作为实验对象.提出了一种基于时间滑动窗口模型的轨迹相似性匹配算法，通过调整时间窗口和位置距离优化算法F值，实现不同网络平台用户的相似性分析.以对新浪微博和滴滴打车的用户位置数据为例进行验证，实验结果证明了地理位置为虚拟身份相似性判断的正相关影响因子，且判断相似性的平均F值超过90%.
|
| 关 键 词： | 社交网络 虚拟身份 位置轨迹 时间滑动窗口 相似性 |
| 收稿时间： | 2021/3/8 0:00:00 |
A Trajectory Similarity Analysis Method Based on Time Sliding Window Model |
| |
| HAN Yi,DU Yanhui,CHEN Qinggang,LU Tianliang.A Trajectory Similarity Analysis Method Based on Time Sliding Window Model[J].Journal of Beijing Institute of Technology(Natural Science Edition),2021,41(11):1207-1214. |
| |
| Authors: | HAN Yi DU Yanhui CHEN Qinggang LU Tianliang |
| |
| Affiliation: | School of Information Network Security, People's Public Security University of China, Beijing 100038, China |
| |
| Abstract: | To solve the problem of difficulty in obtaining personal privacy information of Internet users, incomplete and uncommon or even false content of the public information, a trajectory similarity analysis method was proposed based on time sliding window model. Firstly, some location data with universal, objective and reflecting capability for user behavioral habits were selected as similarity analysis base to collect the data of Sina Weibo and Didi Taxi, forming two high-value data sets with the characteristics of domestic netizens as experimental objects. And then, a trajectory similarity matching algorithm was developed based on the time sliding window. Adjusting the time window and the F value of the location distance optimization algorithm, it was arranged to realize the similarity analysis of users of different network platforms. Finally, some validation experiments were carried out based on the user location data of Sina Weibo and Didi Taxi. The results show that geographical location is a positive correlation factor for virtual identity similarity judgment, and the similarity average F-value can reach up to 90%. |
| |
| Keywords: | social network virtual identity position trajectory time sliding window similarity |
| 本文献已被 万方数据 等数据库收录！ |
| 点击此处可从《北京理工大学学报》浏览原始摘要信息 |
|
点击此处可从《北京理工大学学报》下载全文 |
