面向大规模海量数据的数据挖掘隐私保护方法研究

Research on Privacy Protection Methods for Data Mining in Large-scale and Massive Data

物联网环境产生大量数据,数据隐私保护问题已经成为热点研究领域之一。通过阐述物联网大数据的特点和隐私威胁,分析了现有的数据挖掘隐私保护方法的不足,针对性地提供了一种基于混合高斯分布的数据扰动隐私保护方法。该方法通过生成并公开一组与原始数据独立同分布的新数据的手段来达到对原始数据进行扰动的目的,不仅有效地保护了原始数据隐私,并且保持了原始数据的统计特点,与原始数据上生成的挖掘模型具有相近的准确性。

The Internet of Things environment generates a large amount of data,and the issue of data privacy protection has become one of the hot research fields.By describing he characteristics and privacy threats of big data in the Internet of Things,the shortcomings of existing privacy-preserving methods for data mining are analyzed,a privacy-preserving data perturbation method based on mixed Gaussian distribution is proposed.This method aims to achieve the purpose of perturbing the original data by generating and publicly releasing a set of new data that is independent and identically distributed as the original data.It effectively protects the privacy of the original data while maintaining the statistical characteristics of the original data,and has similar accuracy as the mining model generated on the original data.