INVITED SPEAKERS

• Speaker: Vicenç Torra (IIIA-CSIC, Catalonia) [WEB]
• Institution: IIIA, Artificial Intelligence Research Institute - CSIC, Spanish National Research Council [WEB]
• Title: A new framework to automate constrained microaggregation
• Abstract: Data protection methods from privacy preserving data mining and statistical disclosure control can introduce perturbation in the data. While this perturbation helps to protect the privacy of the respondents, it can introduce inconsistencies and errors. Moreover, the data are normally edited after collection to ensure its correctness and fix inconsistencies, and perturbation methods can introduce new errors in such data. In this paper we present a framework to automate the protection of data with a perturbative method, microaggregation, while at the same time ensuring that no inconsistencies and errors are introduced. That is, the data are microaggregated preserving a set of given constraints (edit constraints).