Abstract: Information security and privacy disclosure become more and more serious in the applications of Internet of things (IoT), which need perfect information security architecture and privacy protection mechanism. There are so many types of anonymization technologies for privacy preserving, but how to evaluate their performance from a quantitative perspective is an important and meaningful research direction. In this paper, based on the analysis and comparison of distance and similarity, several metrics are surveyed. In the following, we use the divergence function to measure the difference between two probability distributions. In the end, we evaluate our method on the Adult dataset from UCI machine learning repository and the result shows the comparison figure of privacy gain and data utility among three commonlyused anonymization privacy preservation technologies.

Key words: privacy preservation, data utility, privacy gain, similarity, divergence

摘要： 摘要在物联网应用推广过程中，信息安全与隐私泄露问题越发明显，需要一个完善的物联网信息安全和隐私保护机制.保护物联网数据隐私的匿名化技术种类较多，如何从定量的角度对不同种类的匿名保护技术的性能进行评估是一个重要且有意义的研究方向.在分析比较了基于距离和基于相似度的度量方法基础上，采用信息论中散度的概念来衡量概率分布间的差异.最后，使用加州大学欧文分校(UCI)机器学习数据库中的数据验证所提方法的可行性，实验结果给出3种常用的匿名隐私保护技术在隐私增益与数据可用性关系的对比.

关键词: 隐私保护, 数据可用性, 隐私增益, 相似度, 散度