Mathematics
Spectral Decomposition
Quick fact
Spectral decomposition is the mathematical equivalent of decomposing a musical chord into its individual notes, each with its own pitch (eigenvalue) and resonance pattern (eigenvector).
Why this is interesting
You’ve seen how a prism splits white light into a rainbow of colors. What if you could do the same for a matrix—breaking it apart into its pure, fundamental frequencies?
Read the full explanation
Understanding Spectral Decomposition
Imagine a transformation that stretches, rotates, or flips a space. Spectral decomposition finds the few special directions (eigenvectors) that the transformation only scales, without rotating. The scaling factors are the eigenvalues. For a symmetric matrix, these directions are perpendicular (orthogonal), like clean axes in a coordinate system. The decomposition expresses the original transformation as a sum of simple projections onto each of these axes, scaled by the corresponding eigenvalue. This is like describing a complex shape as the sum of its shadows on perpendicular walls.
A deeper explanation
The mechanism relies on the fact that a linear operator (e.g., a symmetric matrix) can be diagonalized by an orthogonal basis of eigenvectors. Formally, for a real symmetric matrix A, there exists an orthogonal matrix Q and a diagonal matrix Λ such that A = QΛQᵀ. Each column of Q is an eigenvector, and the diagonal entries of Λ are eigenvalues. This factorization is the spectral decomposition. It matters because it turns complicated matrix powers, inverses, and functions into simple operations on the eigenvalues. For example, computing Aⁿ becomes QΛⁿQᵀ—just raise each eigenvalue to the power. Applications range from Google's PageRank algorithm (finding dominant eigenvectors) to quantum mechanics (observables as operators) and image compression (principal component analysis).