基于频繁活动点集的手机通话位置数据隐私保护方法

项目名称： 基于频繁活动点集的手机通话位置数据隐私保护方法

项目编号： No.41301440

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 天文学、地球科学

项目作者： 尹凌

作者单位： 中国科学院深圳先进技术研究院

项目金额： 25万元

中文摘要： 手机定位数据能够有效的支撑大规模居民活动模式研究。然而，该轨迹数据中蕴含的时空信息可能会造成个人隐私泄露。如何跨越数据挖掘需求和隐私安全性之间的鸿沟进行有益的研究和应用，是当前时空数据挖掘的一个关键研究内容。针对现有轨迹隐私保护方法在频繁活动点集推理攻击下存在的不足,本项目以大规模手机通话位置数据为数据集，完善基于频繁活动点集的隐私模型；以特定的居民活动模式分析为数据使用效益，研究融合某种数据使用效益的高风险活动点集保护方法；并利用国内大城市的大规模手机通话位置数据进行实证分析和方法验证。本项目可丰富大规模轨迹数据发布中的隐私保护理论与方法，促进轨迹数据挖掘的良性可持续发展，有利于突破城市和交通规划中长期缺乏大规模个体活动数据源的限制，并为隐私保护政策和法规制定提供案例支撑和理论依据。

中文关键词： 时空轨迹分析；知识发现；活动空间；位置隐私；

英文摘要： Mobile phone tracking data is able to support studies of individual activity patterns. The rich spatial-temporal information within such trajectory data however may result in privacy breach. There is a problem of finding an optimal trade-off between two conflicting goals: from one side, we would like to have precise data for analytic purposes; from the other side, we would like to have imprecise data to mitigate risks of privacy breach. Dealing with the conflict therefore becomes a critical research question in spatial-temporal data mining. The exsiting privacy preservation approches in the publication of trajectories are limited under inference attacks based on frequent activity locations. Therefore, using a large-scale Call Detail Record as a dataset, this project improves the individual re-identification privacy model based on frequent activity locations. Then this project studies approaches to protecting frequent activity locations with high risks of re-identification, while keeping data utility for certain individual activity analysis. Finally this project uses a large-scale Call Detail Record of a major domestic city as a case study to demonstrate and evaluate the proposed models and approaches. This project will contribute to improve privacy preservation theories and approaches in the publication of traje

英文关键词： trajectory analysis；knowledge discovery；activity space；location privacy；