Comparison between historical population archives and decentralized databases

Differences between large-scale historical population archives and small decentralized databases can be used to improve data quality and record connectedness in both types of databases. A parser is developed to account for differences in syntax and data representation models. A matching procedure is described to discover records from different databases referring to the same historical event. The problem of verification without reliable benchmark data is addressed by matching on a subset of record attributes and measuring support for the match using a different subset of attributes. An application of the matching procedure for comparison of family trees is discussed. A visualization tool is described to present an interactive overview of comparison results.