Table 1. Complete set of 14 pairwise comparison features used by the Random Forest classifiers. Sex encoding and age category features were introduced during Stage 3 (Advanced Feature Engineering).

#	Feature	Type	Description
1	edit_dist_lastname	Continuous [0,1]	Normalized Levenshtein distance of last name
2	edit_dist_middlename	Continuous [0,1]	Normalized Levenshtein distance of middle name
3	edit_dist_phone	Continuous [0,1]	Normalized Levenshtein distance of phone number
4	edit_dist_address	Continuous [0,1]	Normalized Levenshtein distance of street address
5	edit_dist_city	Continuous [0,1]	Normalized Levenshtein distance of city
6	edit_dist_zip	Continuous [0,1]	Normalized Levenshtein distance of ZIP code
7	age_difference	Continuous	Absolute difference in years between dates of birth
8	ssn_match	Binary	1 if Social Security Numbers agree, else 0
9	state_match	Binary	1 if state of residence agrees, else 0
10	record1_sex	Binary	Biological sex of record 1 (Male=0, Female=1)
11	record2_sex	Binary	Biological sex of record 2 (Male=0, Female=1)
12	sex_difference	Binary	1 if the two records differ in sex, else 0
13	record1_agecategory	Categorical	Age category of record 1 (Child / Adult / Older Adult)
14	record2_agecategory	Categorical	Age category of record 2 (Child / Adult / Older Adult)