Conceptual
Browse Instruct Curate
Login

Latent Semantic Analysis in Natural Language Processing

A dimensionality-reduction technique that applies singular value decomposition to a document-term matrix, projecting documents and terms into a lower-dimensional latent space that captures underlying semantic relationships. Used to uncover topics, measure document similarity, and improve information retrieval by grouping terms that co-occur in similar contexts.