Cholesky Decomposition

Let's say AA is symmetric positive definite matrix, then

A=LLTA = LL^T

Where LLis a lower triangular matrix.

Note the cholesky decomposition is unique i.e matrix LLis unique for a given AA.

How to use Cholesky Decomposition

The Cholesky decomposition is mainly used for the numerical solution of linear equations \mathbf {Ax} =\mathbf {b}. If A is symmetric and positive definite, then we can solve\mathbf {Ax} =\mathbf {b} by first computing the Cholesky decomposition {\displaystyle \mathbf {A} =\mathbf {LL} ^{\mathrm {*} }}, then solving {\displaystyle \mathbf {Ly} =\mathbf {b} } for y by forward substitution, and finally solving {\displaystyle \mathbf {L^{*}x} =\mathbf {y} } for x by back substitution.

Square Root

For A=LLTA = LL^T, LLcould be thought of as square root of matrix AA, though mind that it's not a unique square root, many other matrix can be also be thought of as square root of AA.

Application of Cholesky decomposition

Σ=LLT\Sigma = LL^T

Geometrically, the Cholesky matrix transforms uncorrelated variables into variables whose variances and covariances are given by Σ. In particular, if you generate p standard normal variates, the Cholesky transformation maps the variables into variables for the multivariate normal distribution with covariance matrix Σ and centered at the origin.

Last updated