Page cover

Eigen Decomposition

Eigen Decomposition is extremely useful for square symmetric matrices.

The physical interpretation can be given as: "Every real matrix can be thought of as a combination of rotation and stretching."

Av=wARn×nvRn×1wRn×1A\overrightarrow{v} = \overrightarrow{w} \\ A \in \mathbb{R}^{n \times n} \hspace{0.5cm} \overrightarrow{v} \in \mathbb{R}^{n \times 1} \hspace{0.5cm} \overrightarrow{w} \in \mathbb{R}^{n \times 1}

Eigen Vectors

Eigen Vectors of a matrix are those special vectors that only stretch under the action of a matrix.

Eigen Values

Eigen Values are the factor by which eigen vectors stretch.

Av=λvAv = \lambda v

Eigen Decomposition Derivation​

Eigen Decomposition is that set of vectors which only stretch under the action of matrix AA​. Let's say, A(Rn×n)A (\mathbb{R}^{n \times n})​ has nn linearly independent eigenvectors:

{v1,v2,v3,...,vn}\begin{Bmatrix} v^1, v^2, v^3, ..., v^n\\ \end{Bmatrix}

​Concatenate all the vectors (as columns) and make a eigen vector matrix VV​.

V=[v1v2...vn]V = \begin{bmatrix} v^1 & v^2 & ... & v^n\\ \end{bmatrix}

Then, we concatenate the corresponding eigen values into a diagonal matrix:

Λ=diag(λ1,λ2,...,λn)\Lambda = diag(\lambda_1, \lambda_2, ..., \lambda_n)

Eigen Decomposition ​(Factorisation) of AA​ is given by the following equation:

A=VΛV1A = V\Lambda V^{-1}

For real symmetric matrices, we have real eigen vectors and values:

A=QΛQTwhereQT=Q1A = Q \Lambda Q^T \hspace{0.3cm} where \hspace{0.1cm} Q^T = Q^{-1}

Thus, QQ is orthogonal. This also means QQ​ is a rotation matrix. So, for the following action:

Av=QΛQTvAv = Q \Lambda Q^Tv

QTvQ^Tv​ rotates the vector vv​ in the direction of the eigen vector. Then, Λ(QTv)\Lambda(Q^Tv)​ stretches the vector vv​. Finally, Q(ΛQTv)Q (\Lambda Q^Tv)​ rotates vv​ back to its original direction.

Note: Eigen Decomposition may not be unique.

Last updated