一种个性化(p,k)匿名隐私保护算法

现有匿名算法多数仅针对准标识符进行泛化实现隐私保护,未考虑敏感属性的个性化保护问题。为此,在p-sensitive k匿名模型的基础上设计敏感属性个性化隐私保护算法。根据用户自身的敏感程度定义敏感属性的敏感等级,利用敏感属性泛化树发布精度较低的敏感属性值,从而实现对敏感属性的个性化保护。实验结果表明,该算法可有效缩短执行时间,减少信息损失量,同时满足敏感属性个性化保护的要求。     

面向查询服务的数据隐私保护算法

个性化信息服务提高了Web查询精度,但同时也带来数据隐私保护的问题.尤其在面向服务的架构(SOA)中,部署个性化应用时,如何解决隐私保护,这对于个性化服务是一个挑战.随着隐私安全成为微数据发布过程中越来越重要的问题,好的匿名化算法就显得尤为重要.论文总结了前人研究中考虑到准标识符对敏感属性影响的k-匿名算法,提出了直接通过匿名化数据计算准标识符对敏感属性效用的方法以及改进的效用矩阵,同时为了更好地衡量匿名化数据的信息损失,论文中提出了改进的归一确定性惩罚的评价指标,从匿名化数据隐私安全的角度进行分析,实现了改进L-diversity算法,即基于信息损失惩罚的满足L-diversity的算法.它是准标识符对不同敏感属性效用的、并具有较好隐私安全的改进算法.     

基于聚类的高效(K,L)-匿名隐私保护

为防止发布数据中敏感信息泄露,提出一种基于聚类的匿名保护算法.分析易被忽略的准标识符对敏感属性的影响,利用改进的K-means聚类算法对数据进行敏感属性聚类,使类内数据更相似.考虑等价类内敏感属性的多样性,对待发布表使用(K,L)-匿名算法进行聚类.实验结果表明,与传统K-匿名算法相比,该算法在实现隐私保护的同时,数据信息损失较少,执行时间较短.     

基于有损连接的个性化隐私保护

数据表的k-匿名化(k-anonymization)是数据发布环境下保护数据隐私的一种重要方法,在此基础上提出的(,)-匿名模型则是有效的个性化隐私保护方法,泛化/隐匿是实现匿名化的传统技术,然而该技术存在效率低,信息损失量大等缺陷。针对上述问题,引入有损连接的思想,提出了基于贪心策略的(,)-匿名聚类算法,该方法通过准标识符属性和敏感属性间的有损连接来保护隐私数据。实验结果表明,与泛化/隐匿方法相比,该方法在信息损失量和时间效率上具有明显的优势,可以获得更好的隐私信息保护。     

大数据模式分解的隐私保护研究

现有的大多数隐私保护技术往往忽略了敏感属性不同取值和准标识符属性之间存在的特殊关联,并且各领域对数据隐私保护的多方面要求,使得发布的匿名数据需要满足复合隐私约束。对近似敏感属性值和复合隐私约束进行分析,提出了基于大数据模式分解和聚类分析的隐私保护算法。给出了聚类敏感属性值保护相似值方法,设置不同权重的敏感属性,保留重要的属性。使用三维不规则结构矩阵的效用矩阵,来获取精度较高的匿名数据,实现匿名数据的模式分解。在真实数据集上的大量实验结果表明,该算法的数据精确率、数据纠错率都有明显提升,近似攻击率降低。     

基于条件函数依赖的隐私保护模型

数据拥有者发布的数据中如果包含条件函数依赖会导致数据的隐私受到攻击,由条件函数依赖产生的属性间的关联会带来潜在的隐私泄露问题。针对现有的隐私保护方法均无法保护包含条件函数依赖的数据的隐私,形式化地定义了基于条件函数依赖的隐私攻击,提出了隐私保护模型l-deduction来对包含条件函数依赖的数据进行隐私保护;并设计了相应的匿名算法来实现l-deduction模型。理论分析和实验结果表明,该方法既能保护包含条件函数依赖的数据的隐私,又具有较小的信息损失度。     

MA-Datafly:一种支持多属性泛化的k-匿名方法

Datafly算法是数据发布环境下保护数据隐私的一种k-匿名方法,实现k-匿名时只对准标识符属性集中属性值种类最多的属性进行归纳。当准标识符属性集中只有一个属性的取值多样而其他属性取值具有同质性时,该算法可行。实际应用中数据的取值却往往不具有这种特点。针对这个问题,提出一种自底向上的支持多属性归纳k-匿名算法,并对该算法进行实验测试,结果表明该算法能有效降低原始数据的信息损失并能提高匿名化处理效率。     

基于聚类的分级匿名方法

为了防止链接攻击导致隐私的泄露,同时尽可能降低匿名保护时的信息损失,提出(λα, k)-分级匿名模型。该模型根据隐私保护的需求程度,将各敏感属性值划分为高、中、低三个等级类,通过隐私保护度参数λ灵活控制泄露风险。在此基础上,给出一种基于聚类的分级匿名方法。该方法采用一种新层次聚类算法,并针对准标识符中数值型属性与分类型属性采用灵活的概化策略。实验结果显示,该方法能够满足敏感属性的分级匿名保护需求,同时有效地减少信息损失。     

k-匿名数据中的数据依赖问题研究

k-匿名隐私保护模型在隐私保护过程中会产生大量k-匿名数据.为研究k-匿名数据中的数据依赖问题,提出一种扩展函数依赖,将经典函数依赖中的被决定属性取值相等这个条件进行扩展,使其取值来自于同一个指定集合.应用结果表明,该扩展函数依赖不仅包括经典函数依赖、垂直函数依赖、水平函数依赖、度量函数依赖的特性,而且可以从数据完整性的角度描述k-匿名数据的约束条件及指导k-匿名隐私保护模型中准标识符的选取.     

k-匿名方法中准标识符的求解算法

k-匿名是保证视图安全的一种主要手段,如何找出正确的准标识符对k-匿名方法的有效性具有重要意义.针对这一问题,分析了不存在函数依赖和存在函数依赖两种情况下准标识符的组成特征,即当不存在函数依赖时,准标识符由视图间的公共属性组成,当存在函数依赖时,准标识符由视图间的公共属性和秘密信息包含的函数依赖关系的前件属性组成,在此基础上,给出了准标识符的通用求解算法,并用实验证明了算法的有效性和正确性.     

Dimensionality of information disclosure behavior

In studies of people's privacy behavior, the extent of disclosure of personal information is typically measured as a summed total or a ratio of disclosure. In this paper, we evaluate three information disclosure datasets using a six-step statistical analysis, and show that people's disclosure behaviors are rather multidimensional: participants' disclosure of personal information breaks down into a number of distinct factors. Moreover, people can be classified along these dimensions into groups with different “disclosure styles”. This difference is not merely in degree, but rather also in kind: one group may for instance disclose location-related but not interest-related items, whereas another group may behave exactly the other way around. We also found other significant differences between these groups, in terms of privacy attitudes, behaviors, and demographic characteristics. These might for instance allow an online system to classify its users into their respective privacy group, and to adapt its privacy practices to the disclosure style of this group. We discuss how our results provide relevant insights for a more user-centric approach to privacy and, more generally, advance our understanding of online privacy behavior.     

基于差分隐私保护的谱聚类算法

针对传统的聚类算法存在隐私泄露的风险,提出一种基于差分隐私保护的谱聚类算法。该算法基于差分隐私模型,利用累计分布函数生成满足拉普拉斯分布的随机噪声,将该噪声添加到经过谱聚类算法计算的样本相似度的函数中,干扰样本个体之间的权重值,实现样本个体间的信息隐藏以达到隐私保护的目的。通过UCI数据集上的仿真实验,表明该算法能够在一定的信息损失度范围内实现有效的数据聚类,也可以对聚类数据进行保护。     

基于非敏感信息分析的轨迹数据隐私保护发布

针对轨迹数据发布时轨迹和非敏感信息引起的隐私泄露问题,提出一种基于非敏感信息分析的轨迹数据隐私保护发布算法。首先,分析轨迹和非敏感信息的关联性构建轨迹隐私泄露判定模型,得到最小违反序列元组（MVS）,然后借鉴公共子序列的思想,在消除MVS带来的隐私泄露风险时,选择MVS中对轨迹数据损失最小的时序序列作为抑制对象,从而生成具有隐私能力和低数据损失率的匿名轨迹数据集。仿真实验结果表明,与LKC-Local算法和Trad-Local算法相比,在序列长度为3的情况下,该算法平均实例损失率分别降低了6%和30%,平均最大频繁序列（MFS）损失率分别降低了7%和60%,因此所提算法能够有效用于提高推荐服务质量。     

基于小波变换的分布式隐私保护聚类算法

针对现有隐私保护聚类算法无法满足效率与隐私之间较好折中的问题,提出一种基于安全多方计算(SMC)与数据扰动相结合的分布式隐私保护聚类算法。各数据方用小波变换实现数据压缩和信息隐藏,并用属性列的随机重排来防止数据重构可能产生的信息泄露。该算法仅使用压缩重排后的数据参与分布聚类计算,因此计算量和通信量小,算法效率高,而多重保护措施有效保护了隐私数据。因小波变换具有高保真性,所以聚类精度受小波变换的影响较小。理论分析和实验结果表明,所提算法安全高效,在处理高维数据时全局F测量值和执行效率优于基于Haar小波的离散余弦变换(DCT-H)算法,解决了效率与隐私之间的折中问题。     

基于信息浓缩的隐私保护支持向量机分类算法

支持向量机(SVM)的分类决策过程涉及到对原始训练样本的学习,容易导致数据中隐私信息的泄漏。为解决上述问题,提出一种基于信息浓缩的隐私保护分类方法IC-SVM。该算法首先根据样本的邻域信息,通过模糊C均值(FCM)聚类算法进行聚类分析;接着,使用信息浓缩准则对聚类中心进行处理,得到浓缩点组成的新样本;最后,使用新样本进行训练并得到决策函数,并用它去进行分类测试,可以较好地保护数据的隐私。在UCI真实数据和PIE人脸数据上的实验结果表明,IC-SVM方法既能保护数据信息的安全,又有较高的分类准确率。     

防止暴露位置攻击的轨迹隐私保护

为解决移动对象轨迹信息被大量收集所导致的轨迹隐私泄露问题,提出了基于假轨迹的轨迹隐私保护算法。在该算法中,考虑了用户的暴露位置,基于轨迹相似性和位置多样性的综合度量,设计了一种启发式规则来选择假轨迹,从而使得生成的假轨迹能有效隐匿真实轨迹和敏感位置。此外,还提出了轨迹有向图策略和基于网格划分的地图策略来优化算法的执行效率。基于真实的轨迹数据进行实验测试和分析,实验结果表明所提算法在保持数据可用性的情况下能有效保护真实轨迹。     

Preventing sensitive relationships disclosure for better social media preservation

A fundamental aspect of all social networks is information sharing. It is one of the most common forms of online interaction that is tightly associated with social media preservation and information disclosure. As such, information sharing is commonly viewed as a key enabler for social media preservation tasks. In the current situation, where information sharing and inter-user communications are made instantly possible via the widespread use of ubiquitous technologies, privacy related, and particularly information disclosure issues, are the obvious, much discussed, immediate consequences of information sharing. As a result, information disclosure, especially when multimedia data come to play, is critical for appropriate social media preservation strategies that consider and respect the privacy of social network users. Social media preservation must align with privacy protection solutions and consequently must protect sensitive information that social network users would like to keep private. In this paper, we propose a new approach to implement a privacy-oriented social media preservation strategy that prevents the disclosure of sensitive information. Instead of using a preserve-all strategy, we present a framework to personalize social media preservation tasks. We then describe our proposed rule-based algorithm to evaluate information disclosure addressing mainly relationship type disclosure and using shared photos. We also provide an experimental study to investigate the efficiency and the relevance of our approach.     

Factors mediating disclosure in social network sites

In this paper, we explore how privacy settings and privacy policy consumption (reading the privacy policy) affect the relationship between privacy attitudes and disclosure behaviors. We present results from a survey completed by 122 users of Facebook regarding their information disclosure practices and their attitudes about privacy. Based on our data, we develop and evaluate a model for understanding factors that affect how privacy attitudes influence disclosure and discuss implications for social network sites. Our analysis shows that the relationship between privacy attitudes and certain types of disclosures (those furthering contact) are controlled by privacy policy consumption and privacy behaviors. This provides evidence that social network sites could help mitigate concerns about disclosure by providing transparent privacy policies and privacy controls.     

Toward inference attacks for k-anonymity

Current research still cannot effectively prevent an inference attacker from inferring privacy information for k-anonymous data sets. To solve the issue, we must first study all kinds of aggressive reasoning behaviors and process for the attacker thoroughly. Our work focuses on describing comprehensively the inference attack and analyzing their privacy disclosures for k-anonymous data sets. In this paper, we build up a privacy inference graph based on attack graph theory, which is an extension of attack graph. The privacy inference graph describes comprehensively the inference attack in k-anonymous databases by considering attacker background knowledge and external factors. In the privacy inference graph, we introduce a concept of valid inference path to analyze the privacy disclosures in face of inference attack. According to both above, we design an algorithm to compute the n-valid inference paths. These paths can deduce some privacy information resulting in privacy disclosure. Moreover, we study the optimal privacy strategies to resist inference attack by key attribute sets and valid inference paths in the attack graph. An approximate algorithm is designed to obtain the approximate optimal privacy strategy set. At last, we prove the correctness in theory and analyze the performance of the approximate algorithm and their time complexity.     

Efficient and privacy-preserving skyline computation framework across domains

Skyline computation, which returns a set of interesting points from a potentially huge data space, has attracted considerable interest in big data era. However, the flourish of skyline computation still faces many challenges including information security and privacy-preserving concerns. In this paper, we propose a new efficient and privacy-preserving skyline computation framework across multiple domains, called EPSC. Within EPSC framework, a skyline result from multiple service providers will be securely computed to provide better services for the client. Meanwhile, minimum privacy disclosure will be elicited from one service provider to another during skyline computation. Specifically, to leverage the service provider’s privacy disclosure and achieve almost real-time skyline processing and transmission, we introduce an efficient secure vector comparison protocol (ESVC) to construct EPSC, which is exclusively based on two novel techniques: fast secure permutation protocol (FSPP) and fast secure integer comparison protocol (FSIC). Both protocols allow multiple service providers to calculate skyline result interactively in a privacy-preserving way. Detailed security analysis shows that the proposed EPSC framework can achieve multi-domain skyline computation without leaking sensitive information to each other. In addition, performance evaluations via extensive simulations also demonstrate the EPSC’s efficiency in terms of providing skyline computation and transmission while minimizing the privacy disclosure across different domains.