Singular value decomposition
Singular value decomposition: A matrix factorisation into orthogonal components and singular values describing principal scales and directions. SVD supports compression, denoising and rank analysis. Truncation preserves dominant components but discards potentially useful information. This perspective connects the definition with practical behaviour in artificial intelligence systems.
In italiano: Decomposizione ai Valori SingolariLegge I · Embedding, feature e spazi vettorialiSVD supports compression, denoising and rank analysis. Truncation preserves dominant components but discards potentially useful information.