Technical GlossaryNatural Language Processing

Stemming

An approach that reduces a word to a shorter root-like form by crudely stripping suffixes.

Stemming is used to reduce word variation, especially in information retrieval and classical text mining. However, it does not guarantee linguistic correctness, and the resulting form may not correspond to a true lexical root. Even so, it can help reduce dimensionality and improve matching coverage. In agglutinative languages such as Turkish, it should be used carefully and often alongside more context-aware alternatives.