作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程

• 安全技术 • 上一篇    下一篇

移动轨迹数据去匿名化攻击方法

钟建友,常姗,刘晓强,宋晖   

  1. (东华大学 计算机科学与技术学院,上海 201620)
  • 收稿日期:2015-12-07 出版日期:2016-12-15 发布日期:2016-12-15
  • 作者简介:钟建友(1989—),男,硕士研究生,主研方向为移动网络隐私保护;常姗(通讯作者),副教授、博士;刘晓强、宋晖,教授、博士。
  • 基金资助:
    国家自然科学基金(61300199,61402101);中央高校基本科研业务费专项资金(2232014D3-21,2232014D3-42);上海自然科学基金(14ZR1400900)。

De-anonymization Attack Method for Mobile Trace Data

ZHONG Jianyou,CHANG Shan,LIU Xiaoqiang,SONG Hui   

  1. (School of Computer Science and Technology,Donghua University,Shanghai 201620,China)
  • Received:2015-12-07 Online:2016-12-15 Published:2016-12-15

摘要: 为保护移动对象轨迹隐私,轨迹数据集发布前常使用假名对轨迹进行匿名化处理。然而,假名用户的匿名轨迹仍面临隐私泄露风险。为此,提出一种新的去匿名化攻击方法。攻击者若获得其攻击对象当前或未来任意时段的若干轨迹片段,则可以此比对匿名历史轨迹数据集,从中识别出攻击对象的历史轨迹。对2组真实移动轨迹数据进行特征分析,给出基于轨迹特征相似度的去匿名方法。采用改进的词频-逆文档频率方法提取历史轨迹的特征向量,通过主成分分析降维后,对历史轨迹和攻击者所获得的轨迹片段进行特征匹配,识别出与攻击者所持有轨迹特征相似度最高的历史轨迹。实验结果表明,所提方法可获得较高的去匿名准确率。

关键词: 移动轨迹, 假名, 轨迹隐私, 去匿名化, 特征提取

Abstract: To protect the trace privacy of mobile objects,pseudonym is used to the anonymous processing of trace before the release of the trace dataset.However,the anonymous trace of pseudonym users still faces the risk of privacy leakage.This paper proposes a new de-anonymization attack method.If an attacker obtains several track segments of his attack target at present or in any future period,comparing the traces with the anonymous historical trace dataset,the historical traces of the attack target are identified.The characteristics of the real moving track data of the two groups are analyzed,and a de-anonymization method based on characteristic similarity is presented.The feature vectors of history trace are extracted based on improved Term Frequency-Inverse Document Frequency(TF-IDF) method.The dimension is reduced by Principal Component Analysis(PCA),and the feature matching is performed on the track segments obtained by the historical track and the attacker,to recognize the historical trace with the highest degree of similarity with the trace characteristics of attackers.Experimental results show that the proposed method can obtain higher accuracy.

Key words: mobile trace, pseudonym, trace privacy, de-anonymization, feature extraction

中图分类号: