JOVANA
Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

Symmetric Matrices and the Spectral Theorem

Symmetric matrices (A = A^T) are the nicest matrices in all of linear algebra. The [[spectral-theorem|spectral theorem]] guarantees they have real [[eigenvalue|eigenvalues]] and a full [[orthonormal-basis|orthonormal]] set of eigenvectors, so they factor cleanly as A = Q*D*Q^T with [[orthogonal-matrix|Q orthogonal]]. We also meet [[positive-definite-matrix|positive definite]] matrices along the way.

What makes a matrix symmetric

A matrix is symmetric when it equals its own transpose: A = A^T. The entry in row i, column j equals the entry in row j, column i. These show up everywhere real data lives — covariance matrices, distance and similarity tables, the matrices behind graphs and physical energy.

Symmetry is not a cosmetic detail. It forces the matrix to behave beautifully, and the theorem that says exactly how is the spectral theorem.

The spectral theorem

The spectral theorem says: every real symmetric matrix has all real eigenvalues, and you can always find a full set of eigenvectors that are mutually orthogonal and unit length — an orthonormal basis. No complex numbers, no missing directions, no awkward repeated cases. Compare this to general matrices, which may have complex eigenvalues or fail to have enough eigenvectors at all.

A = Q*D*Q^T
  Q : columns are orthonormal eigenvectors (Q^T = Q^(-1))
  D : diagonal matrix of the real eigenvalues
Because Q is orthogonal, its inverse is just its transpose — the cleanest diagonalization there is.

Positive definite: when every eigenvalue is positive

A symmetric matrix is positive definite when all its eigenvalues are strictly positive. Geometrically every direction gets stretched, never flipped, so the quadratic form x^T*A*x is always positive for x not zero — a bowl that curves up in every direction. Covariance matrices, and the matrices in many optimization and physics problems, are positive definite.