We know that a positive definite matrix has a Cholesky decomposition,but I want to know how a Cholesky decomposition can be done for positive semi-definite matrices?The following sentences come from a paper. "There are two assumptions on the specified correlation matrix R. The first is a general assumption that R is a possible correlation matrix, i.e. that it is a symmetric positive semidefinite matrix with 1’s on the main diagonal. While implementing the algorithm there is no need to check positive semi-definiteness directly, as we do a Cholesky decomposition of the matrix R at the very start. If R is not positive semi-definite, the Cholesky decomposition will fail." Thank you for your answer.
$\begingroup$
$\endgroup$
8
-
$\begingroup$ THANK YOU .but here I want to know if a positive semi-definite can be done for Cholesky decomposition? and how? $\endgroup$– PurpleCommented Jan 20, 2014 at 10:39
-
1$\begingroup$ "that how a positive semi-definite be done for Cholesky decomposition"??? What do you mean? $\endgroup$– Dima PasechnikCommented Jan 20, 2014 at 11:28
-
$\begingroup$ excuse for my bad english."There are two assumptions on the specified correlation matrix R. The first is a general assumption that R is a possible correlation matrix, i.e. that it is a symmetric positive semidefinite matrix with 1’s on the main diagonal. While implementing the algorithm there is no need to check positive semi-definiteness directly, as we do a Cholesky decomposition of the matrix R at the very start. If R is not positive semi-definite, the Cholesky decomposition will fail." $\endgroup$– PurpleCommented Jan 20, 2014 at 11:46
-
4$\begingroup$ en.m.wikipedia.org/wiki/Cholesky_decomposition has a proof that yes, indeed, it is correct, a p.s.d. R will have a Cholesky decomposition. $\endgroup$– Dima PasechnikCommented Jan 22, 2014 at 20:08
-
1$\begingroup$ Thank you for your answer.In Matlab Programming, the "chol" command can be used to simply apply this to only a positive definite matrix.When it comes to a positive semi-definite matrix，“chol” fails.I want to ask if there any command in Matlab Programming that can be used to a positive semi-definite matrix? $\endgroup$– PurpleCommented Jan 23, 2014 at 7:11
|
Show 3 more comments
2 Answers
$\begingroup$
$\endgroup$
You can either:
use a LDL^T decomposition (see e.g. here)
deflate the kernel yourself before: that is, compute a basis $Q_2$ for the kernel, complete it to a square orthonormal matrix $Q=[Q_1 \, Q_2]$, and assemble $$ Q^TRQ=\begin{bmatrix}R_{11} & 0\\ 0 & 0\end{bmatrix}, $$ where $R_{11}$ is going to be nonsingular (and hence can be Cholesky-factored).
answered Jan 20, 2014 at 10:25
$\begingroup$
$\endgroup$
If you don't mind including some permutations, you can get a variant of Cholesky that still has the rank-revealing property:
$$P^T R P = R_1^T R, \quad R_1 = \begin{bmatrix} R_{11} & R_{22} \\ 0 & 0 \end{bmatrix}.$$
This is a matter of simple greedy pivoting. For the algorithm and more details, see Higham's "Cholesky Factorization".
answered Aug 28, 2016 at 12:58