On the connections between statistical disclosure control for microdata and some artificial intelligence tools

Abstract

Statistical disclosure control (SDC) and artificial intelligence (AI) use similar tools for different purposes. This work describes the common elements of both areas to increase their synergy. SDC is a discipline that seeks to modify statistical data so that they can be published (typically by National Statistical Offices) without giving away the identity of any individual behind the data. When dealing with individual data (microdata in SDC jargon), both SDC procedures and AI knowledge integration procedures use similar principles for different purposes (masking data vs. improving its quality). Similarities can also be found for methods evaluating re-identification risk in SDC and data mining tools for making data consistent. This paper explores those methodological connections with the aim of stimulating interaction between both fields. In particular, data mining turns out to be a common interest of both fields.