基于k-对称匿名算法的社会网络隐私保护研究

发布时间：2018-05-14 09:31

本文选题：社会网络 + 隐私保护　；参考：《河南大学》2014年硕士论文

【摘要】：近年来，随着互联网技术的不断发展，社交网络产品也在不断的融入我们的生活中。从QQ、人人网到微博、微信，，社交网络渐渐成为我们生活中不可或缺的一部分。但是，社交网络在提供给我们便利的同时也对我们个人隐私保护及社会关系隐私保护提出了新的挑战。现阶段在传统的关系型数据库隐私保护研究领域已经有了很多科研成果，但是由于社会网络的数据模型是类似于计算机图论中图的结构，因此我们在处理社会网络隐私保护问题时显然不能直接套用针对传统关系型数据库的隐私保护方法。然而，伴随着大数据时代的到来，我们一般处理的社会网络数据也是海量的，人工处理显然不现实，因此社会网络的隐私保护问题必是当前研究的热点问题也是未来计算机技术必然的研究趋势。目前我们使用社会网络软件主要是为了与他人共享或者交换信息资源，单纯的个人信息隐私保护已经不能满足需求，对个人社会关系隐私保护的研究是目前的热门研究领域。本文主要从数据挖掘的角度对k-匿名算法进行研究。首先介绍了现阶段社会网络隐私保护研究的国内外现状及其概念和特点，针对性的分析了攻击社会网络的几种方式，并对现阶段几种匿名算法进行了介绍。以此为基础，借鉴他人已有的研究思想，对原k-对称匿名算法给予改进，并设计出一种有效地还原算法，找出一个推导出k值的公式。k-对称匿名方法是一种隐私保护算法，对社会网络中的节点进行对称处理，使得等价类的结果中每个集合都包括k个节点，这就使得攻击者识别目标个体的概率不高于1/k。还针对k-对称匿名方法的可用性分析提出一种能还原出原社会网络图的还原算法。最后，论文基于微信讨论组的社会网络数据，实现了k-对称匿名发布，评估了这种匿名发布方法的可用性，并且验证了有效性。
[Abstract]:In recent years, with the continuous development of Internet technology, social network products are constantly integrated into our lives. From QQ, Renren to Weibo, WeChat, social networks are becoming an integral part of our lives. However, social networks not only provide us with convenience, but also pose new challenges to our privacy protection and social privacy protection. At present, a lot of achievements have been made in the field of privacy protection of traditional relational database, but the data model of social network is similar to the structure of graph in computer graph theory. Therefore, when we deal with the social network privacy protection problem, we obviously can not directly apply the traditional relational database privacy protection method. However, with the arrival of the big data era, the social network data we generally deal with is also massive, and manual processing is obviously not realistic. Therefore, the privacy protection of social networks must be the hot topic of current research and the inevitable research trend of computer technology in the future. At present, we use social network software mainly to share or exchange information resources with others. The simple privacy protection of personal information can no longer meet the needs. The research on privacy protection of personal social relations is a hot research field at present. In this paper, the k-anonymity algorithm is studied from the point of data mining. Firstly, this paper introduces the present situation, concepts and characteristics of the research on privacy protection of social networks at home and abroad, analyzes several ways of attacking social networks, and introduces several anonymous algorithms at the present stage. On this basis, the original k-symmetric anonymous algorithm is improved, and an effective algorithm is designed to find a formula to deduce k value. The method of k-symmetric anonymity is a privacy protection algorithm. The nodes in the social network are treated symmetrically so that each set of the equivalent class includes k nodes which makes the probability of the attacker to identify the target individual not more than 1 / 1 / k. Based on the availability analysis of k-symmetric anonymous method, an algorithm is proposed to restore the original social network graph. Finally, based on the social network data of WeChat discussion Group, the paper implements k-symmetric anonymous publishing, evaluates the availability of this anonymous publishing method, and verifies its effectiveness.
【学位授予单位】：河南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.08

【参考文献】