Record linkage is used to establish relationships between records of two different data files. In this work, record linkage is studied for files that correspond to the same set of individuals but that do not share a common set of variables. Under this circumstance, classical techniques can not be applied. We present an approach to this problem based on clustering techniques and knowledge integration ones. In this way, common underlying structures in both files can be detected and re-identification is possible. This approach is based on some basic assumptions that are made explicit in this work.
Keywords: Data Mining and Knowledge Discovery
Citation: VicenÁ Torra: Towards the Re-identification of Individuals in Data Files with Non-common Variables. In W.Horn (ed.): ECAI2000, Proceedings of the 14th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2000, pp.326-330.