Method for Noise Addition for Individual Record Preserving Privacy and Statistical Characteristics: Case Study of Real Estate Transaction Data.
Yuzo Maruyama, Ryoko Tone, Yasushi Asami · arXiv (Cornell University) · 2015
Abstract: We propose a new method to perturb a major variable by adding noise, so that the result of regression analysis is not affected by the noise addition. The extent of the perturbation can be controlled by a single parameter, which will ease the perturbation operation in the actual application. Through a numerical experiment, we recommend an appropri-ate value of the parameter to achieve both the sufficient perturbation to mask the original values and the sufficient coherence between the perturbed data and the original data. 1.