摘要： 数据挖掘技术具有很多优点，但存在隐私威胁等不足。该文针对聚类分析时如何保护隐私的问题，提出独立噪音思想并设计独立噪音算法(INA)。该算法对原数据叠加噪音以保护原始数据不被泄漏，所用噪音不会对数据分布造成严重影响，使后期挖掘工作可以在修改后的数据上直接进行。实验结果证明，INA算法可以取得较高的隐私保护程度和挖掘正确率。

关键词: 隐私保护, 聚类挖掘, 独立噪音

Abstract: Data mining technique has a lot of merits, but it faces criticism like privacy threatening. This paper focuses on how to preserve privacy in clustering analysis and introduces the Independent Noise Algorithm(INA). This algorithm preserves original data by adding noises while keeping its distribution. The noises won’t influence the data distributing badly so that the mining can be done directly on the amended data later. The experimental results demonstrate that INA is effective and provide acceptable values for balancing privacy and accuracy.

Key words: privacy preserving, clustering mining, independent noise

中图分类号: 

• TP309.2