Computer-based genealogy reconstruction in founder populations

This paper describes a software tool that reconstructs entire genealogies from data collected from different and heterogeneous sources, including municipal and parish records archived over centuries. The tool exploits a record linkage algorithm relying on a rule-based data matching approach. It applies a general strategy for managing the ambiguities due to missing, imprecise or erroneous input data. The process follows an iterative approach that combines automatic pedigree reconstruction with software-empowered human data revision to improve the quality and the accuracy of the results and to optimize the matching rules.The paper discusses the results obtained by reconstructing the entire genealogy of the population of the Val Borbera, a geographically isolated valley in Northern Italy. The genealogy could be reconstructed from data going back as far as the XVI century. The resulting pedigree includes 75,994 trios, 58.9% of which belonging to a unique big family, reconstructed over 13 generations.

Figure optionsDownload as PowerPoint slideHighlights
► A software tool to reconstruct entire genealogies from data collected over centuries.
► It exploits record linkage techniques relying on a rule-based data matching strategy.
► It combines automatic reconstruction with software-empowered human data revision.
► We discuss the results on the entire genealogy of the Val Borbera population.
► Final pedigree: 75.994 trios, 58,9% belonging to a unique family of 13 generations.