An eigenvector keeps its direction under A and is only scaled: Av = λv.
6 min read508 wordscore5 viewsupdated 2026-10-03
Definition
An eigenvector of a square matrix A is a nonzero vector v that A maps onto its own line: Av=λv. The scalar λ is the corresponding eigenvalue — the factor by which that direction is stretched (negative λ flips it).
At a glance
Notation
Av=λv; A=VΛV−1 with eigenvectors as columns of V and Λ=diag(λ1,…,λn)
Input → output
a square n×n matrix → n eigenvalues (with multiplicity, possibly complex) and their eigenvectors
Key property
trA=∑iλi, detA=∏iλi; symmetric A has real λ and orthonormal eigenvectors
Prerequisites
matrix multiplication, determinant, linear systems & rank
Used in
PCA, PageRank, spectral clustering, Markov chains, curvature of loss surfaces
The intuition
A matrix moves almost every vector off its own line: it turns the vector as well as stretching it. Eigenvectors are the exceptions — the directions the map leaves pointing the same way. Along them, a complicated matrix acts like a plain number.
That is why they matter. If you write any vector in eigenvector coordinates, applying A just multiplies each coordinate by its own λ, and applying A a hundred times multiplies by λ100. Directions with ∣λ∣>1 blow up, those with ∣λ∣<1 die out, and the largest ∣λ∣ eventually dominates everything.
The math
Av=λv means (A−λI)v=0 has a nonzero solution, i.e. A−λI is singular. So the eigenvalues are the roots of the characteristic polynomial, and each eigenvector spans part of the null space of A−λI:
det(A−λI)=0,(A−λI)v=0,v=0
I
identity matrix of the same size as A
λ
an eigenvalue — a root of the degree-n polynomial
For a 2×2 matrix the polynomial is λ2−(trA)λ+detA, so
λ1,2=2trA±4(trA)2−detA
trA
trace, a11+a22
detA
a11a22−a12a21
A negative discriminant gives a complex pair: a rotation has no real direction it preserves. A zero discriminant gives a repeated eigenvalue, which may come with only one eigenvector — the shear (1011) has λ=1 twice but only the direction (1,0).
Diagonalization. If A has n independent eigenvectors, stack them as the columns of V: then AV=VΛ, so A=VΛV−1 and powers become cheap, Ak=VΛkV−1. For symmetric A the spectral theorem guarantees real eigenvalues and an orthonormal V=Q:
A=QΛQ⊤=i=1∑nλiqiqi⊤,λmax=∥x∥=1maxx⊤Ax
qi
orthonormal eigenvectors (columns of Q)
x⊤Ax
Rayleigh quotient of a unit vector x
Power iteration. Write x0=∑icivi. Then Akx0=∑iciλikvi=λ1k(c1v1+∑i>1ci(λi/λ1)kvi). If ∣λ1∣>∣λ2∣ and c1=0, repeatedly multiplying by A and normalizing converges to ±v1, and the tangent of the angle error shrinks by ∣λ2/λ1∣ every step.
Power iteration
1Start from a random unit vector x (random ensures c1=0).
fig.x → Ax: the symmetric preset [[2, 1], [1, 2]] stretches the ring into an ellipse; only the highlighted directions (0.71, 0.71) and (0.71, −0.71) stay on their dashed lines, scaled by λ₁ = 3 and λ₂ = 1 (trace 4, det 3). The rotation preset gives λ = 0.80 ± 0.60i and no real direction; shear has a single one. Power iteration on [[2, 0.5], [0.5, 1]] (λ₁ = 2.207, λ₂ = 0.793) starts 75.6° away and is within 3.72° after four steps — the tangent of the error falls by λ₂/λ₁ = 0.359 per step.
Worked example
A=(2112): trA=4, detA=3, so λ2−4λ+3=(λ−3)(λ−1)=0 and λ1=3, λ2=1. For λ1=3, A−3I=(−111−1) kills v1=(1,1); check A(1,1)=(3,3). For λ2=1, v2=(1,−1); check A(1,−1)=(1,−1). The eigenvectors are orthogonal, as the spectral theorem promises.
Powers: A10=Qdiag(310,1)Q⊤=21(59050590485904859050)=(29525295242952429525) — no repeated multiplication needed.
Power iteration from x0=(1,0) (unnormalized): (2,1)→(5,4)→(14,13)→(41,40). The angle to (1,1) is 18.43°→6.34°→2.12°→0.71°; its tangent goes 31,91,271,811, shrinking by exactly λ2/λ1=1/3 each step.
In code
python
1import numpy as np
2
3A = np.array([[2.0, 1.0], [1.0, 2.0]])
4lam, V = np.linalg.eig(A) # lam ≈ [3., 1.]; columns of V are unit eigenvectors
5print(np.allclose(A @ V, V * lam)) # True: A v_i = λ_i v_i, column by column
eig for general matrices, eigh for symmetric ones (real, sorted, orthonormal), and a hand-written power iteration.
Where it's used
PCA
Principal components are the eigenvectors of the covariance matrix; each λi is the variance captured along qi.
PageRank & Markov chains
The stationary distribution is the eigenvector of the transition matrix with λ=1, found by power iteration over the web graph.
Spectral clustering
The eigenvectors of a graph Laplacian with the smallest eigenvalues embed nodes so that k-means can separate weakly connected groups.
Optimization curvature
Hessian eigenvalues give curvature along each direction: gradient descent on a quadratic diverges unless the step size is below 2/λmax, and mixed signs mark a saddle point.
Check yourself
What are the eigenvalues of A=(4213)?
Check yourself
A symmetric matrix has eigenvalues λ1=2 and λ2=1.8 (all others smaller). By what factor does power iteration shrink the tangent of its angle error per step?
Remember
✓Av=λv: eigenvectors are the directions A only scales; λ is the scale.
✓Eigenvalues solve det(A−λI)=0; ∑λi=trA and ∏λi=detA.
✓Symmetric matrices factor as QΛQ⊤ with real λ and orthonormal Q — the basis of PCA.
✓Power iteration finds the dominant eigenvector at rate ∣λ2/λ1∣ per step.