Table 1.

Description of the data used for evaluating the configuration optimization approach.

Data set	Description	Original records, n	Duplicates, n
FEBRL1a	Distributed with the FEBRL package.	500	500 (1 per original)
FEBRL2	Distributed with the FEBRL package.	4000	1000 (maximum 5 per original)
FEBRL3	Distributed with the FEBRL package.	2000	3000 (maximum 5 per original)
FEBRL4	Distributed with the FEBRL package.	5000	5000 (1 per original)
Hawaii	Constructed with the FEBRL data set generator using a number of Hawaii-specific data sources.	1000	1000 (maximum 5 per original)

^aFEBRL: Freely Extensible Biomedical Record Linkage.