RELAIS: An Open Source Toolkit for Record Linkage

The combined use of statistical and administrative sources allow to save time and money, reducing survey costs, response burden, etc.; sometimes data sources are hard to integrate since errors or lacking information in the record identifiers may complicate this process. The purpose of record linkage is to identify the same real world entity, which can be differently represented in data sources. To deal with record linkage complexity and application dependency, we propose a toolkit called RELAIS (REcord Linkage At IStat). The toolkit is based on the idea of choosing the most appropriate technique for each phase, and of dynamically combining such techniques in order to build a workflow, on the basis of application constraints and data features provided as input. RELAIS is configured as an open source project giving the possibility of gathering together the efforts already done in the scientific community towards the definition of a record linkage project. A real case study validates the RELAIS idea.