EIGEN-THEORY

Eigen (German for "Own"). The characteristic vectors that define the hidden axes of rotation and absolute stability of a mathematical system.

01 // The Golden Rule

Normally, when you multiply a vector by a matrix, it changes direction entirely. But an Eigenvector is special: it refuses to turn. It is the invisible spine of the transformation. It only gets stretched or squashed.

A\vec{v} = \lambda\vec{v}

Matrix TransformScalar Scaling

Alignment Scanner

Matrix A

\begin{bmatrix} 2 & 1 \\ 1 & 2 \end{bmatrix}

Input Angle (θ)0°

Drag to find the invariant axes.

Status

Vectors Unaligned

Input

\vec{v}

Output

A\vec{v}

02 // The Anatomy

The Direction

Eigenvector

(\vec{v})

The axis of the transformation. Think of it as the axle of a spinning wheel—it is part of the wheel, but it remains perfectly stationary while everything else rotates around it.

The Magnitude

Eigenvalue

(\lambda)

The stretching factor. If $\lambda = 2$ , the eigenvector doubles in length. If $\lambda = -1$ , the space flips along that axis entirely.

Decoding the Matrix

The Characteristic Equation

To hunt down the eigenvalues, we must solve a specialized determinant equation. We rearrange $A\vec{v} = \lambda\vec{v}$ to mathematically ask: "By how much must we shift the main diagonal so that the matrix flattens space to zero?"

\det(A - \lambda I) = 0

A

Matrix

\lambda I

Shift

04 // Why We Care

PageRank

Google's entire original algorithm is based on finding the dominant eigenvector of the web's link graph.

Resonance

Bridge collapses (like Tacoma Narrows) happen when external wind frequencies perfectly match the structure's eigenvalues.

Quantum Mechanics

The famous Schrödinger equation is fundamentally just an eigenvalue problem determining energy states.

Facial Recognition

Algorithms use 'Eigenfaces' to dramatically compress and simplify image data for rapid identification.

Invariant Axes Located

You are ready to compress data streams with SVD.

Next: SVD