Improved Assessment of the Accuracy of Record Linkage via an Extended MaCSim Approach

Abstract Record linkage is the process of bringing together same entity from overlapping data sources while removing duplicates. Huge amounts are now being collected by public or private organizations as well researchers and individuals. Linking analysing relevant information this massive reservoir can provide new insights into society. It has become increasingly important to have effective efficient methods for linking different sources. Therefore, it becomes necessary assess ability a method achieve high accuracy compare between with respect accuracy. In article, we improve on Markov Chain based Monte Carlo simulation approach ( MaCSim ) assessing method. The improvement proposed here involves calculation similarity weight every variable value each record pair, which allows partial agreement values. To method, correctly linked proportions investigated record. extended illustrated using synthetic set provided Australian Bureau Statistics realistic settings. Test results show assessment linkages.