Overview

In linear algebra, an eigenvector of a linear transformation is a nonzero vector that does not change its direction when the transformation is applied, although its length may be scaled or it may be reversed. The scale factor associated with an eigenvector is called an eigenvalue (commonly denoted by the Greek letter λ). Formally, for a linear operator represented by a matrix A, an eigenvector v and eigenvalue λ satisfy a relation of the form A v = λ v. Eigenpairs (λ, v) summarize how a transformation acts on special directions in the space.

Key properties

Eigenvalues and eigenvectors capture structural features of matrices and linear maps. Important points include:

  • Direction invariance: eigenvectors maintain their line through the origin under the map, possibly reversing orientation when λ is negative.
  • Zero eigenvalue: λ = 0 indicates the transformation maps the eigenvector to the zero or null vector, and signals non-invertibility of the operator.
  • Multiplicity: an eigenvalue can have an algebraic multiplicity (root multiplicity of a characteristic equation) and a geometric multiplicity (dimension of its eigenspace); these may differ.
  • Complex values: even for real matrices, eigenvalues and eigenvectors can be complex, reflecting rotations and combinations of stretching.

History and terminology

The adjective "eigen" comes from German, meaning "own", "intrinsic", or "characteristic". The concepts emerged during 19th-century studies of quadratic forms, vibrations, and differential equations, where identifying natural modes and frequencies required analysis of special vectors and scalars associated with linear operators.

Computation and examples

Analytically, eigenvalues are found by solving the characteristic equation det(A − λI) = 0; corresponding eigenvectors lie in the nullspace of A − λI. For small matrices this can be done by hand; for larger matrices numerical algorithms are used. Common methods include the power method for dominant eigenvalues and more robust procedures such as the QR algorithm. In practice, eigenvectors of a covariance matrix give principal components in statistics, while eigenmodes describe natural vibration shapes in mechanics.

Applications and distinctions

Eigenanalysis appears across mathematics, science and engineering: stability analysis of dynamical systems, principal component analysis in data science, quantum mechanics where operators have eigenstates and eigenvalues, and graph theory where eigenvalues of adjacency or Laplacian matrices encode connectivity properties. It is useful to contrast eigenvectors with singular vectors from the singular value decomposition (SVD): singular vectors relate to orthogonal decompositions and nonnegative singular values, while eigenvectors arise directly from square operators and can involve complex values. For further background on linear maps and function types see related topics.

For a concise introduction that balances theory and computation, consult standard linear algebra texts or reliable online resources. Practical computation typically relies on established numerical libraries that implement the algorithms described above.