2008.50: A Preconditioned Newton Algorithm for the Nearest Correlation Matrix
2008.50: Rudiger Borsdorf and Nicholas J. Higham (2008) A Preconditioned Newton Algorithm for the Nearest Correlation Matrix. IMA Journal of Numerical Analysis, 30 (1). pp. 94-107. ISSN 1464-3642
This is the latest version of this eprint.
Full text available as:
|PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader|
Various methods have been developed for computing the correlation matrix nearest in the Frobenius norm to a given matrix. We focus on a quadratically convergent Newton algorithm recently derived by Qi and Sun. Various improvements to the efficiency and reliability of the algorithm are introduced. Several of these relate to the linear algebra: the Newton equations are solved by minres instead of the conjugate gradient method, as it more quickly satisfies the inexact Newton condition; we apply a Jacobi preconditioner, which can be computed efficiently even though the coefficient matrix is not explicitly available; an efficient choice of eigensolver is identified; and a final scaling step is introduced to ensure that the returned matrix has unit diagonal. Potential difficulties caused by rounding errors in the Armijo line search are avoided by altering the step selection strategy. These and other improvements lead to a significant speedup over the original algorithm and allow the solution of problems of dimension a few thousand in a few tens of minutes.
Advance Access published on June 8, 2009.
|Uncontrolled Keywords:||correlation matrix, positive semidefinite matrix, Newton's method, preconditioning, rounding error, Armijo line search conditions, alternating projections method|
|Subjects:||MSC 2000 > 65 Numerical analysis|
MSC 2000 > 90 Operations research, mathematical programming
|Deposited By:||Nick Higham|
|Deposited On:||28 January 2010|
Available Versions of this Item
- A Preconditioned Newton Algorithm for the Nearest Correlation Matrix (deposited 28 January 2010) [Currently Displayed]