基于位置的社会网络隐私保护关键技术

发布时间：2018-10-04 18:47

【摘要】：随着无线传感器网络和定位技术的蓬勃发展,越来越多的用户开始使用基于位置的社会网络(LBSN)应用,由此产生了大量的LBSN数据。这些数据在挖掘用户的行为习惯、挖掘用户之间的相似性、为用户推荐朋友或感兴趣的地点等方面,起到了重要作用。然而直接发布LBSN数据会给用户带来隐私泄露问题,当前还没有针对LBSN数据的匿名方法,因此本文重点研究面向LBSN数据的隐私保护技术。目前虽然存在大量的社会网络隐私保护技术和位置匿名技术,但是这些技术都不适用于LBSN数据。LBSN数据既包含了社会网络中的用户关系,又包含了用户与位置之间的关系,使得LBSN数据的匿名技术更具有挑战性。本文首先对社会网络隐私保护技术和位置匿名技术进行了系统综述,针对攻击者可能得到的不同背景知识,提出了两种LBSN图上的两种不同的隐私泄露问题,包括结点身份泄露和敏感关系泄露问题,和用户结点rank序列导致的结点身份泄露问题。为了解决结点身份泄露和敏感关系泄露问题,本文提出了k-连接安全分组匿名模型,并分别提出了k--LSG算法和k-ILSG算法实现该匿名模型。k-LSG算法先对位置结点直接进行k-安全分组,然后处理分组中产生的外连分组和内连分组,使它们都满足k-连接安全分组条件。进而提出了支持高可用性的k-ILSG算法,以提高匿名数据可用性。其次,为了解决LBSN图中用户结点的rank序列导致的结点身份泄露问题,本文提出了LBSN图的K-rank匿名模型,并提出了启发式算法RA来得到LBSN图的匿名rank序列,以及根据匿名rank序列重构出匿名LBSN图的RRL算法。最后,通过在真实数据集上的大量实验,分别验证了本文所提出的LBSN图的k-连接安全分组匿名方法和K-rank匿名方法的安全性以及匿名数据的高可用性。
[Abstract]:With the rapid development of wireless sensor networks and localization technology, more and more users begin to use location-based social network (LBSN) applications, which results in a large amount of LBSN data. These data play an important role in mining the behavior habits of users, mining the similarity between users, recommending friends or places of interest to users, and so on. However, publishing LBSN data directly will bring privacy disclosure to users. There is no anonymous method for LBSN data at present. Therefore, this paper focuses on privacy protection technology for LBSN data. Although there are a lot of social network privacy protection techniques and location anonymity techniques, these technologies are not suitable for LBSN data. LBSN data not only includes the user relationship in social network, but also includes the relationship between user and location. This makes the anonymity of LBSN data more challenging. In this paper, firstly, the privacy protection technology and location anonymity technology of social network are systematically reviewed. According to the different background knowledge that the attacker may obtain, two different kinds of privacy disclosure problems on two kinds of LBSN diagrams are proposed. It includes node identity leakage and sensitive relation leakage, and node identity leakage caused by user node rank sequence. In order to solve the problem of node identity leakage and sensitive relationship disclosure, this paper proposes a k- join secure packet anonymous model, and proposes the k--LSG algorithm and k-ILSG algorithm to implement the anonymous model. The k-LSG algorithm implements k- security grouping directly to the location node first. Then, the external and internal packets generated in the packets are processed to satisfy the security packet condition of k- connection. Furthermore, a high availability k-ILSG algorithm is proposed to improve the availability of anonymous data. Secondly, in order to solve the problem of node identity leakage caused by the rank sequence of user nodes in LBSN graph, this paper proposes the K-rank anonymous model of LBSN graph, and proposes a heuristic algorithm RA to obtain the anonymous rank sequence of LBSN graph. And the RRL algorithm based on anonymous rank sequence to reconstruct anonymous LBSN graph. Finally, through a large number of experiments on real data sets, the security of the K-Connected secure grouping Anonymous method and K-rank Anonymous method and the high availability of Anonymous data are verified respectively in this paper.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.5;TP212.9

【相似文献】