Free to use, no account

SVD Calculator

Decompose a small rectangular matrix A into compact factors U, Σ and Vᵀ. See singular values, estimated numerical rank, the reconstructed matrix and its error, then download the results as CSV.

Use 2–8 rows and 2–8 columns. Separate values with spaces or commas; put each row on a new line. Decimal and scientific notation are supported. Values must be between −1,000,000 and 1,000,000.

Numerical floating-point estimate, not an exact symbolic result. Rank near the tolerance can vary; the compact factors are intended for small matrices.

How do I calculate an SVD?

Enter a rectangular matrix with one row per line and values separated by spaces or commas. The calculator estimates A = UΣVᵀ and checks how closely the factors reconstruct your matrix.

Frequently asked questions

How do I calculate an SVD?

Enter a rectangular matrix with one row per line and values separated by spaces or commas. The calculator estimates A = UΣVᵀ and checks how closely the factors reconstruct your matrix.

Is my matrix uploaded?

No. The decomposition runs locally in your browser; the matrix is not uploaded or saved by this tool.

What is singular value decomposition?

SVD factors a matrix as A = UΣVᵀ, where the singular values in Σ describe the strength of orthogonal directions.

Can this calculator handle rectangular matrices?

Yes. Enter between 2 and 8 rows and columns; the returned factors use a compact rectangular form.

Is the numerical rank exact?

No. It is estimated using a floating-point tolerance, so very small singular values may be treated as zero.