A posteriori disclosure risk measure for tabular data based on conditional entropy

Abstract

Statistical database protection, also known as Statistical Disclosure Control (SDC), is a part of information security which tries to prevent published statistical information (tables, individual records)from disclosing the contribution of specific respondents. This paper deals with the assessment of the disclosure risk associated to the release of tabular data. So-called sensitivity rules are currently being used to measure the disclosure risk for tables. These rules operate on an a priori basis: the data are examined and the rules are used to decide whether the data can be released as they stand or should rather be protected. In this paper, we propose to complement a priori risk assessment with a posteriori risk assessment in order to achieve a higher level of security, that is, we propose to take the protected information into account when measuring the disclosure risk. The proposed a posteriori disclosure risk measure is compatible with a broad class of disclosure protection methods and can be extended for computing disclosure risk for a set of linked tables. In the case of linked table protection via cell suppression, the proposed measure allows detection of secondary suppression patterns which offer more protection than others.