For a linear transformation ; a non-zero vector is an eigenvector of if:
for some scalar , which is called the eigenvalue associated with eigenvector .
We can compute numerically:
- In NumPy, we use
np.linalg.eig
, which outputs a named tuple, with attributes eigenvalues and eigenvectors. - In MATLAB, we use
[V,D] = eig(A)
, which outputs a diagonal matrix of eigenvaluesD
and a matrix with corresponding eigenvectorsV
.
Definitions
We also define a basis as an eigenbasis for if the vectors are eigenvectors of . A handy tool is the characteristic polynomial, where is an eigenvalue of if and only if .
Theorems
- Eigenvalues which correspond to distinct eigenvectors are linearly independent.
- If has distinct eigenvalues, then is diagonalisable.
- The trace of is the sum of its eigenvalues, i.e.:
Computing the eigenvalues/eigenvectors of a matrix is doable with diagonalisation. Note that any real symmetric matrix can be decomposed into:
where is composed of the eigenvectors of , and is a diagonal matrix.
Multiplicity
We define two helpful metrics:
Algebraic multiplicity: how many times an eigenvalue occurs, denoted . Geometric multiplicity: the dimension of the eigenspace , denoted .
We also observe by theorem, .