Eigenvalues and Eigenvectors

College Depth 58 in the knowledge graph I know this Set as goal
Unlocks 4093 downstream topics
eigenvalues eigenvectors invariant

Core Idea

For matrix A, scalar λ is an eigenvalue if Av = λv has a nonzero solution v (an eigenvector). Eigenvectors span the eigenspace E_λ = nul(A − λI). Eigenvalues reveal how A stretches directions; they are invariant under similarity transformations and determine the dynamics of matrix iteration.

Explainer

Most linear transformations scramble vectors — they change both direction and length. But every square matrix has special directions it simply scales, never rotates. These are the eigenvectors. If Av = λv for a nonzero vector v, then multiplying v by A just multiplies it by the scalar λ (the eigenvalue). The vector "survives" the transformation unchanged in direction.

You already know how to find eigenvalues from the characteristic polynomial: solve det(A − λI) = 0. Each root λ is an eigenvalue. Once you have λ, you find its eigenvectors by solving (A − λI)v = 0 — that is, finding the null space of (A − λI). This null space is the eigenspace E_λ. It can be one-dimensional (a line of eigenvectors through the origin) or higher-dimensional if multiple independent vectors satisfy the equation.

Eigenvalues tell you how the matrix stretches along each eigendirection. A positive eigenvalue λ > 1 stretches; 0 < λ < 1 compresses; λ < 0 flips direction and scales; λ = 0 collapses the eigenvector to zero, which signals that A is singular. In applied settings — Google's PageRank, principal component analysis, quantum mechanics, and population dynamics — the largest eigenvalue and its eigenvector describe the dominant long-run behavior of a system under repeated matrix application.

One subtlety: eigenvalues are determined by the characteristic polynomial (so they are fixed properties of A), but eigenvectors are not unique. Any nonzero scalar multiple of an eigenvector is also an eigenvector with the same eigenvalue. The eigenspace captures the entire family. This is why we talk about eigenspaces rather than individual eigenvectors — the direction matters, not the specific vector.

Eigenvalues and eigenvectors unlock diagonalization: if A has n linearly independent eigenvectors, it can be written as A = PDP⁻¹ where D is diagonal with eigenvalues on the diagonal. This makes repeated matrix multiplication trivial (Aⁿ = PDⁿP⁻¹) and is the foundation for efficient algorithms in differential equations, Markov chains, and matrix exponentiation.

Practice Questions 3 questions

Prerequisite Chain

Counting to 10Counting to 20Understanding ZeroThe Number ZeroCounting to FiveOne-to-One CorrespondenceCombining Small Groups Within 5Addition Within 10Addition Within 20Two-Digit Addition Without RegroupingTwo-Digit Addition with RegroupingAddition Within 100Repeated Addition as MultiplicationMultiplication Facts Within 100Division as Equal SharingDivision as Grouping (Measurement Division)Division: Grouping (Repeated Subtraction) ModelDivision: Fair Sharing ModelDivision as Equal SharingDivision as GroupingBasic Division FactsDivision Facts Within 100Two-Digit by One-Digit DivisionDivision with RemaindersRemainders and Quotients in DivisionDivision Word ProblemsIntroduction to Long DivisionFactors and MultiplesPrime and Composite NumbersEquivalent FractionsRelating Fractions and DecimalsDecimal Place ValueIntegers and the Number LineOpposites and Additive InversesAbsolute ValueAdding IntegersSubtracting IntegersMultiplying IntegersDividing IntegersUnit RatesProportionsPercent ConceptConverting Between Fractions, Decimals, and PercentsOperations with Rational NumbersTwo-Step EquationsSolving Multi-Step EquationsEquations with Variables on Both SidesLiteral EquationsSlope-Intercept FormPoint-Slope FormWriting Linear EquationsParallel and Perpendicular Line SlopesGraphing Linear EquationsSystems of Equations — Graphing MethodSystems of Equations — Elimination MethodSystems of Three VariablesMatrices IntroductionLinear TransformationsEigenvalues and Eigenvectors

Longest path: 59 steps · 234 total prerequisite topics

Prerequisites (2)

Leads To (78)

Adjacency Matrix and Spectral Basicshard Adjacency Matrix and Spectral Graph Theoryhard Baseline New Keynesian Modelhard Capital Asset Pricing Model (CAPM)soft Characteristic Equation and Closed-Loop Stabilityhard Characteristic Polynomial and Diagonalizationhard Chemometrics and Multivariate Data Analysissoft Chemometrics and Multivariate Data Analysissoft Confirmatory Factor Analysis and Measurement Validationsoft Confirmatory Factor Analysis and Measurement Validationhard Controllability and Observabilityhard Convergence of Iterative Methodshard Coupled Oscillator Systems and Equations of Motionhard Crystal Structure and Bravais Latticessoft DSGE Models: Dynamic Stochastic General Equilibriumsoft DSGE Models: Dynamic Stochastic General Equilibriumhard Diagonalizationhard Diagonalization and Similar Matriceshard Dimensionality Reduction Techniquessoft Eigenvalue Method for Systems of ODEshard Eigenvalues and Eigenstateshard Factor Analysis and Dimensionality Reductionhard Factor Analysis and Dimensionality Reductionsoft Factor Analysis and Measurement Modelssoft Factor Analysis and Measurement Modelssoft Factor Models and Multifactor Pricing (Fama-French)soft Fixed Points and Stabilityhard Focal Mechanisms and Stress Tensorshard Hilbert Spaces and Dirac Notationsoft Hydrogen Atom: Quantum Energy Levels and Orbitalshard Kernel Theory and RKHSsoft Linear Regression in Machine Learningsoft Linearization and the Jacobian Matrixhard Machine Learning Applications in Social Sciencehard Machine Learning Applications in Social Sciencesoft Matrix Representationssoft Molecular Orbital Theory: LCAO-MOsoft Multicollinearity: Detection Using VIFhard Multidimensional Item Response Theorysoft Multidimensional Item Response Theoryhard Multilevel Modeling for Hierarchical Datasoft Normal Modes and Collective Oscillationshard Phase Portraits for Linear Systemshard Pole Placement via State Feedback and Observer Designhard Portfolio Diversificationsoft Power Method for Eigenvalueshard Principal Component Analysishard Principal Component Analysissoft Quantum Angular Momentumhard Quantum Fourier Transformsoft Quantum Mechanical Treatment of Hydrogenhard Quantum Operatorshard Quantum Phase Estimationhard Real Business Cycle Theorysoft Reciprocal Lattice and Brillouin Zonessoft Regularization Theory (Tikhonov, Spectral)hard Schur's Lemmasoft Social Network Analysis: Structural Positions and Dynamicssoft Social Network Analysis: Structural Positions and Dynamicssoft Solving the Schrödinger Equation for Hydrogen Atomhard State Feedback and Pole Placementhard State Feedback and Pole Placementhard State Transformations and Similarity Transformationssoft State Transition Matrixhard State-Space Representationsoft State-Space Representationhard Steady-State Analysis in Growth Modelshard Steady-State Growth and Balanced Growth Pathhard Structural Equation Modeling with Latent Variableshard Structural Equation Modeling with Latent Variablessoft Symmetric Matrices and Their Propertiessoft The Schrödinger Equationhard The Schrödinger Equationhard The Variational Principle and Trial Wavefunctionssoft Variational Quantum Eigensolver (VQE)hard Vector Autoregression (VAR) Modelshard Vector Autoregression (VAR) Modelssoft Vector Autoregression (VAR) Models and Impulse Responseshard