At least one implementation, described herein, detects fuzzy duplicates and eliminates such duplicates. Fuzzy duplicates are multiple, seemingly distinct tuples (i.e., records) in a database that represent the same real-world entity or phenomenon.