Linked User Records by Similarity

Problem A data-quality system needs to locate user profiles that are duplicates or sufficiently alike. The input records is a two-dimensional string array. Its first row is a header containing field names, for example "id", "name", "email", and "company"; every later row stores one profile's string values. You also receive weights, whose entries use the format " : ". Every named field must appear in the header and must be different from "id". Each weight is a decimal value…