Singular value decomposition (svd) in MATLAB

Introduction

Singular value decomposition breaks a matrix into simple parts to understand its structure and solve problems easily.

To find the main features or patterns in data, like in image compression.

To solve systems of equations that are hard to solve directly.

To reduce the size of data while keeping important information.

To analyze and clean noisy data by focusing on strong signals.

To compute matrix properties like rank or condition number.

Syntax

MATLAB

[U, S, V] = svd(A)

A is the input matrix you want to decompose.

U and V are orthogonal matrices, and S is a diagonal matrix with singular values.

Examples

Decompose a 2x2 matrix A into U, S, and V matrices.

MATLAB

A = [3 1; 1 3];
[U, S, V] = svd(A);

Decompose a random 4x3 matrix to get its singular values and vectors.

MATLAB

A = rand(4,3);
[U, S, V] = svd(A);

Sample Program

This program decomposes matrix A using svd and shows the three resulting matrices U, S, and V.

MATLAB

A = [4 0 0; 3 -5 0];
[U, S, V] = svd(A);
disp('Matrix U:');
disp(U);
disp('Matrix S:');
disp(S);
disp('Matrix V:');
disp(V);

Important Notes

The singular values in S are always non-negative and sorted from largest to smallest.

U and V matrices are orthogonal, meaning their columns are perpendicular and have length 1.

You can use svd to approximate a matrix by keeping only the largest singular values.

Summary

Singular value decomposition splits a matrix into U, S, and V parts.

It helps understand data structure and solve math problems easily.

Use svd in MATLAB with [U, S, V] = svd(A).