Using Noise Addition Method Based on Pre-mining to Protect Healthcare Privacy

Abstract

With medical device cyber-physical systems being more and more widely used, a lot of healthcare data are produced, making data sharing for health research a vital requirement. But, privacy concerns must be addressed before sharing and publishing any data set. Privacy-preserving data mining (PPDM) is an important technology to protect personal privacy. This paper begins with a proposal of two new noise addition algorithms for perturbing the original healthcare data, and then applies them to a two-step perturbation model. Experiments show that the algorithms given in this paper have much higher accuracy than existing ones under the similar privacy strength.