A workshop at the University of Guelph May 24-25 2010
OVC LifeLong Learning Centre Rm 1713
Record linkage at the Minnesota Population Center
Ron Goeken, Lap Huynh, Tom Lenius, and Rebecca Vick (Minnesota Population Center)
This paper will present an overview of methods used to link various samples of the United States Population Censuses to a complete-count database of the 1880 United States population. Topics include name standardization, construction of similarity scores and the use of support vector machines to classify linked records (SVMs). We discuss our preliminary data release and subsequent work on our final release. Topics include the construction of name commonness scores and birth density measures, and their impact on the final linked data. We also present a number of indirect measures assessing the accuracy of our linked data. We also discuss the construction of weights to deal with linkage rate differentials.
An Automated Record Linkage System – Linking 1871 Canadian Census to 1881 Canadian Census
Luiza Antonie (U of Guelph), Peter Baskerville (U of Alberta), Kris Inwood (U of Guelph) and Andrew Ross (U of Guelph)
This paper describes a recently developed linkage system for historical Canadian censuses and its application for tracking people from 1871 to 1881. The record linkage system incorporates a supervised learning module for classifying pairs of records as matches and non-matches. The classification module was trained using a set of true links that was created by experts. We evaluate the first results and provide a road map for further experimentation.
Using family lineage data to improve record linkage success
David Barss (Family Search)