MERLIN -- A Tool for Multi-party Privacy-Preserving Record Linkage

Abstract

Many organizations, including businesses, government agencies and research organizations, are collecting vast amounts of data, which are stored, processed and analyzed to mine interesting patterns and knowledge to support efficient and quality decision making. In order to improve data quality and to facilitate further analysis, many application domains require information from multiple sources to be integrated and combined. The process of matching and aggregating records that relate to the same entities from different data sources without compromising their privacy is known as 'privacy-preserving record linkage' (PPRL), 'blind data linkage' or 'private record linkage'. In this paper we present MERLIN, an online tool that demonstrates various PPRL methods in a multi-party context. In this demonstration we show different private multi-party blocking and matching techniques, and illustrate the usability of MERLIN by presenting quality and performance measures of various PPRL methods. We believe MERLIN will help practitioners and researchers to better understand the pipeline of the PPRL process, to compare different multi-party PPRL techniques, and to determine the best technique to use for their needs.