An approach to measures of privacy in randomized response models for quantitative characteristics

摘要

In this paper, we present a new and reasonable measure of privacy protection of randomized response with or without allowing for the auxiliary variable and compare the unrelated question model, the addition model, the multiplication model, the optional randomization model, the generalized multiplicative model and the forced model in the light of the efficiency and protection privacy. This article shows that the results are sometimes in contrast with the authors' assertions regarding the superiority of their methods. This measure is easy to evaluate.