SVD
Compute the singular value decomposition (SVD) of any 2x2 real matrix online for free. Factor A into USVT with orthogonal matrices U, V and singular values sigma1, sigma2 for linear algebra and data science.
About This Calculator
This calculator computes the complete singular value decomposition (SVD) of any 2x2 real matrix. Given a matrix A = [[a, b], [c, d]], the calculator factors it into A = USigmaVᵀ, where U and V are orthogonal matrices and Sigma is a diagonal matrix containing the singular values sigma1 >= sigma2 >= 0 on its diagonal.
How SVD is computed: The calculator first computes AᵀA = [[a^2+c^2, ab+cd], [ba+dc, b^2+d^2]] and finds its eigenvalues by solving lambda^2 - tr(AᵀA)lambda + det(AᵀA) = 0. The singular values are sigma1 = sqrtlambda1 and sigma2 = sqrtlambda2. The columns of V are the eigenvectors of AᵀA, normalized to unit length. Finally, U is computed from AV = USigma, giving uᵢ = (1/sigmaᵢ)Avᵢ for each non-zero singular value. The result displays all three matrices: U, Sigma, and Vᵀ, along with the individual singular values sigma1 and sigma2.
Why SVD matters: The singular value decomposition is one of the most powerful tools in linear algebra. It reveals the fundamental structure of a matrix by breaking it down into its constituent geometric transformations: rotation (Vᵀ), scaling (Sigma), and another rotation (U). The singular values sigma1 and sigma2 represent the scaling factors along the principal axes of the transformation. The largest singular value sigma1 equals the spectral norm (operator norm) of the matrix, and the ratio κ = sigma1/sigma2 is the condition number that measures numerical stability.
Applications of SVD: SVD is used extensively in data science for Principal Component Analysis (PCA) and dimensionality reduction. In image processing, the SVD enables lossy image compression by retaining only the largest singular values. Recommendation systems (like those used by Netflix and Amazon) use SVD-based matrix factorization to predict user preferences. In signal processing, SVD helps separate noise from meaningful data. Google's PageRank algorithm relies on similar matrix factorization techniques. In engineering, SVD is used to solve ill-conditioned linear systems and compute the pseudoinverse of non-square matrices.
How to use: Enter the four entries a11, a12, a21, a22 of your 2x2 matrix and click Decompose. The calculator displays the orthogonal matrices U and Vᵀ, the diagonal matrix Sigma with singular values, and the two individual singular values sigma1 and sigma2. Use the share button to copy a unique URL containing your matrix values for future reference or sharing.
Frequently Asked Questions
What is singular value decomposition (SVD)?
Singular value decomposition (SVD) is a matrix factorization that expresses any real matrix A as A = USigmaVᵀ, where U and V are orthogonal matrices and Sigma is a diagonal matrix containing the singular values sigma1 >= sigma2 >= ... >= 0. For a 2x2 matrix, U and V are 2x2 orthogonal matrices and Sigma = [[sigma1, 0], [0, sigma2]].
How do you calculate SVD of a 2x2 matrix by hand?
To compute SVD of a 2x2 matrix by hand: (1) Compute AᵀA and find its eigenvalues lambda1, lambda2. (2) The singular values are sigma1 = sqrtlambda1 and sigma2 = sqrtlambda2. (3) The columns of V are the eigenvectors of AᵀA. (4) Compute U from AV = USigma, where each column uᵢ = (1/sigmaᵢ)Avᵢ for non-zero sigmaᵢ. (5) Arrange U, Sigma, and Vᵀ to verify A = USigmaVᵀ.
What are U, Sigma, and V in SVD?
In SVD factorization A = USigmaVᵀ: U is an mxm orthogonal matrix whose columns are the left singular vectors (eigenvectors of AAᵀ). Sigma is an mxn diagonal matrix with the singular values sigma1 >= sigma2 >= ... >= 0 on its diagonal. V is an nxn orthogonal matrix whose columns are the right singular vectors (eigenvectors of AᵀA).
What are singular values used for in real-world applications?
Singular values and SVD are fundamental in data science and engineering. Applications include: Principal Component Analysis (PCA) for dimensionality reduction, image compression (approximating images by retaining only the largest singular values), recommendation systems (Netflix prize matrix factorization), signal denoising, Google's PageRank algorithm, and solving ill-conditioned least-squares problems in scientific computing.
Is SVD the same as eigenvalue decomposition?
No, SVD and eigenvalue decomposition are different. SVD applies to any matrix (square or rectangular), while eigenvalue decomposition requires a square matrix. SVD always produces real non-negative singular values, while eigenvalues can be negative or complex. For a symmetric positive semi-definite matrix, the SVD coincides with the eigenvalue decomposition: U = V and the singular values equal the absolute eigenvalues.
What does a zero singular value mean in SVD?
A zero singular value means the matrix is rank-deficient (singular). For a 2x2 matrix, if sigma2 = 0, the matrix has rank 1, meaning its columns are linearly dependent. Geometrically, the matrix transformation collapses the 2D space onto a 1D line. The number of non-zero singular values equals the rank of the matrix.
How is the condition number related to singular values?
The condition number of a matrix is κ = sigma1/sigma2, the ratio of the largest to the smallest singular value. A high condition number (κ ≫ 1) indicates an ill-conditioned matrix where small input errors can cause large output errors in linear systems. A condition number close to 1 means the matrix is well-conditioned. This is crucial in numerical analysis and engineering.
Is this tool free to use?
Yes, all calculators on Calculy are completely free to use with no registration required. You can share your SVD results via a unique URL containing your matrix values.