A methodology for multi-source Record Linkage.
As patient's medical data is disseminated in different health structures, building a data-warehousing system at a regional level, for healthcare co-operation, has some specific requirements compared to intra healthcare structure data-warehousing projects. In particular, there is no common patient electronic record's identifier among all the sources. Hence, data integration and Record Linkage is the most important issue in such projects. In this paper, we present Background on Information integration with the two main approaches of the problem: Schema Integration approach and Entity reconciliation approach. We show the limits of each of the two approaches and propose an approach based on theory of a set partitioning and instance identification.