Revisiting distance-based record linkage for privacy-preserving release of statistical datasets

Abstract

Statistical Disclosure Control (SDC, for short) studies the problem of privacy-preserving data publishing in cases where the data is expected to be used for statistical analysis. An original dataset T containing sensitive information is transformed into a sanitized version T which is released to the public. Both utility and privacy aspects are very important in this setting. For utility, T must allow data miners or statisticians to obtain similar results to those which would have been obtained from the original dataset T . For privacy, T must significantly reduce the ability of an adversary to infer sensitive information on the data subjects in T .One of the main a-posteriori measures that the SDC community has considered up to now when analyzing the privacy offered by a given protection method is the Distance-Based Record Linkage (DBRL) risk measure. In this work, we argue that the classical DBRL risk measure is insufficient. For this reason, we introduce the novel Global Distance-Based Record Linkage (GDBRL) risk measure. We claim that this new measure must be evaluated alongside the classical DBRL measure in order to better assess the risk in publishing T instead of T . After that, we describe how this new measure can be computed by the data owner and discuss the scalability of those computations.We conclude by extensive experimentation where we compare the risk assessments offered by our novel measure as well as by the classical one, using well-known SDC protection methods. Those experiments validate our hypothesis that the GDBRL risk measure issues, in many cases, higher risk assessments than the classical DBRL measure. In other words, relying solely on the classical DBRL measure for risk assessment might be misleading, as the true risk may be in fact higher. Hence, we strongly recommend that the SDC community considers the new GDBRL risk measure as an additional measure when analyzing the privacy offered by SDC protection algorithms.

A Unified Framework for Quantifying Privacy Risk in Synthetic Data

“…The concepts of linkability and inference attacks have extensively been studied both on synthetically generated datasets [56] and datasets that were anonymized with other approaches [25,37,59]. While many practical inference attacks exist, e.g.…”

Cloud-Assisted Privacy-Preserving Method for Healthcare Using Adaptive Fractional Brain Storm Integrated Whale Optimization Algorithm

“…Hence, the dyadic product can be replaced with some other advanced concepts for further improvements. (iii) In [3], the selection of applicable trusted infrastructure, service provider, and algorithms is still inadequate to satisfy user confidentiality requirements. (iv) e security method in [4] can be extended by including some effective sensitive attributes from the anonymous data in order to improve security.…”

“…e perturbation approach produces some changes to input data, whereas the generalization approach replaces the original elements with less accurate elements, and synthetic data generators generate the synthetic data similar to the original data [3,4]. Moreover, other protection methods employed to ensure the secrecy of information are data sanitation, blocking, cryptography, and anonymization.…”

The role of statistics in the era of big data: Electronic health records for healthcare research

“…(Harron et al, 2015) A range of sophisticated linkage algorithms exist. For example, EM algorithms using matching weights, (Belin and Rubin, 1995) distance-based algorithms, (Herranz et al, 2015) and prior-informed imputation, (Harron et al, 2014) have all been investigated. Others have explored machine learning methods, which have potential for very large linkage problems (Elfeky et al, 2003) and further guidance on the value of these methods is needed.…”