For an square matrix , its eigenvectors are non-zero vectors such that results in scalar multiple of . In other words, can scale, reverse the direction of, or rotate , but it will not otherwise transform it. can have at most linearly independent eigenvectors.

For each eigenvector , there is a corresponding characteristic (possibly complex-valued) scalar called an eigenvalue .

The eigenvalues of a square matrix can be discovered by calculating the determinant and solving for . The eigenvectors can then be discovered by solving for each . (See Eigenvalue decomposition for a square matrix.)

Interpretation of eigenvalues

For any eigenvalue :

  • The scaling factor is .
  • If has negative sign, the direction of the eigenvector is reversed.

Complex eigenvalues always appear in pairs, as complex conjugates. Such a pair of eigenvalues corresponds to a rotation in a 2-dimensional (sub)space of the -dimensional space. The angle of rotation in this subspace is given by

See also Interpretation of eigenvalues and eigenvectors in ordinary differential equations.