支持轨迹隐私保护的两阶段用户兴趣区构建方法

发布时间：2018-04-11 11:05

本文选题：数据聚类 + 空间数据　；参考：《计算机学报》2017年12期

【摘要】：针对空间大数据开放共享中平衡隐私保护与数据可用性矛盾的需求,该文从空间和时间两个维度对空间时序数据进行分析,提出蕴含空间、时间和群体特征的用户兴趣区构建方法.该方法分为两个阶段:第一阶段的个人兴趣区构建首先将m个移动用户的轨迹数据预处理到n个采样时刻,并形式化为隐含时序关系的m×n阶位置矩阵;然后根据访问频率等指标对每个用户在位置矩阵中的行向量进行聚类、合并和优化,获得每个用户在不同时间段的若干个人兴趣区.第二阶段的公共兴趣区构建在第一阶段的基础上,首先对每个用户按照一定的选取方式提取代表个人兴趣区位置信息的位置点,并对全部m个移动用户的个人兴趣区的位置点进行二次聚类,获得所有用户在不同时间尺度上的若干公共兴趣区;然后根据实际应用场景需要,提取包含时间标记的公共兴趣区.通过对比公开数据集的处理结果和百度地图的实体数据,验证了该方法所构建的用户兴趣区与实际生活中的功能区域基本一致.应用实例证明,该文方法所构建的用户兴趣区可以为空间大数据开放共享中的轨迹隐私保护提供有效的技术支撑.
[Abstract]:In order to balance the contradiction between privacy protection and data availability in the open sharing of space big data, this paper analyzes spatial temporal data from two dimensions of space and time, and puts forward the implied space.Time and group characteristics of user interest area construction method.The method is divided into two stages: in the first stage, the region of personal interest is constructed by preprocessing the trajectory data of m mobile users to n sampling times, and formalizing the m 脳 n order position matrix of implicit temporal relation;Then the row vectors of each user in the location matrix are clustered, merged and optimized according to the access frequency and other indicators, and several areas of personal interest of each user in different time periods are obtained.In the second stage, the public interest area is constructed on the basis of the first stage. Firstly, the location points representing the location information of the area of personal interest are extracted from each user according to a certain selection method.The location points of all m mobile users' personal areas of interest are clustered twice to obtain some common areas of interest of all users on different time scales, and then according to the needs of practical application scenarios,Extract the area of public interest that contains the time tag.By comparing the processing result of the public data set and the entity data of Baidu map, it is verified that the user interest area constructed by this method is basically consistent with the function area in real life.The application examples show that the user interest area constructed by this method can provide an effective technical support for the path privacy protection in the open sharing of space big data.
【作者单位】：西安交通大学电子与信息工程学院;陕西省计算机网络重点实验室;
【基金】：国家自然科学基金(61472316,61502380) 中央高校基本科研业务费综合交叉重点项目(XKJC2014008) 陕西省重大基础研究项目(2016ZDJC-05)资助~~
【分类号】：TP309

【相似文献】