Quantifying Privacy: A Novel Entropy-Based Measure of Disclosure Risk
This addresses the challenge of quantifying privacy for researchers and practitioners using data mining, though it appears incremental as it builds on existing privacy models.
The paper tackles the problem of evaluating and comparing privacy protection techniques by proposing a novel entropy-based measure of disclosure risk, which they empirically apply to methods like query restriction, sampling, and noise addition.
It is well recognised that data mining and statistical analysis pose a serious treat to privacy. This is true for financial, medical, criminal and marketing research. Numerous techniques have been proposed to protect privacy, including restriction and data modification. Recently proposed privacy models such as differential privacy and k-anonymity received a lot of attention and for the latter there are now several improvements of the original scheme, each removing some security shortcomings of the previous one. However, the challenge lies in evaluating and comparing privacy provided by various techniques. In this paper we propose a novel entropy based security measure that can be applied to any generalisation, restriction or data modification technique. We use our measure to empirically evaluate and compare a few popular methods, namely query restriction, sampling and noise addition.