Fuzzy matching is a machine learning problem dating back to the 1980s that involves assessing the similarity of two strings that are not identical. The algorithms quantify the similarity by measuring the “edit distance” between the strings and producing an approximation match based on a similarity score. Fuzzy matching is often used interchangeably with Fuzzy Search, which focuses on information retrieval from a database based on similar but non-exact entries.

This technique operates under fuzzy logic, which indicates the degree to which a statement is true, in contrast to traditional Boolean logic. The main limitation is that it rarely considers the semantics of the text, making it inferior to semantic or entity-based approaches for tasks like internal linking opportunity identification. Use cases include redirect mapping, product name standardization, and analyzing misspelled brand mentions.