Computing trace of function of matrix

MaRDI QIDQ2721776zbMATH OpenFDO

Authors Hongbin Guo

Publication date 10 July 2001

Published in Numerical Mathematics (Search for Journal in Brave)

zbMATH Keywords

algorithmcondition numberGauss type quadrature ruleLanczos procedurelarge scale matricesMonte Carlo methodquadratic formtrace function

Mathematics Subject Classification ID

Monte Carlo methods (65C05) Numerical quadrature and cubature formulas (65D32) Numerical computation of matrix norms, conditioning, scaling (65F35)

Review text

The author gives a further improvement of an algorithm based on a Monte Carlo method and a Gauss type quadrature rule for computing the trace function of large scale matrices. The condition number of the problem is analyzed and the error of the algorithm is estimated. It is shown that the relative error is controllable.

Recommendations

• Fast estimation of \(\mathrm{tr}(f(A))\) via stochastic Lanczos quadrature
• A posteriori error estimate for computing \(\operatorname{tr}(f(A))\) by using the Lanczos method.
• Estimating the condition number of \(f(A)b\)
• Estimates of the trace of the inverse of a symmetric matrix using the modified Chebyshev algorithm
• The extended global Lanczos method for matrix function approximation

Cited in

(3)

• Estimating the trace of the matrix inverse by interpolating from the diagonal of an approximate inverse
• Probing for the Trace Estimation of a Permuted Matrix Inverse Corresponding to a Lattice Displacement
• UPRE method for total variation parameter selection

This page was built for publication: Computing trace of function of matrix

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2721776)