Accurate Density of States Estimation: A Comparative Study

DOI10.48550/arXiv.2609.31953MaRDI QIDQ7360699arXivFDO

Authors Dong Min Roh, Chao Yang

Publication date 25 September 2026

Published in arXiv

Tldr We compare different density of states (DOS) estimators for large sparse symmetric matrices: Gaussian-broadened stochastic Lanczos, Jackson-damped kernel polynomial approximation, monotone piecewise-cubic interpolation of CDOS midpoint data, and Gaussian-process regression with single- and double-Gaussian covariance kernels.

Full work available at URL https://doi.org/10.48550/arXiv.2609.31953

arXiv classification

math.NA

Cites work

• A Method for Constructing Local Monotone Piecewise Cubic Interpolants
• A new metric for probability distributions
• A spectrum adaptive kernel polynomial method
• Analysis of stochastic Lanczos quadrature for spectrum approximation
• Approximating spectral densities of large matrices
• Divergence measures based on the Shannon entropy
• Gaussian Processes for Machine Learning
• Monotone Piecewise Cubic Interpolation
• Numerical methods for large eigenvalue problems
• Polynomial based iteration methods for symmetric linear systems
• Randomized algorithms for estimating the trace of an implicit symmetric positive semi-definite matrix
• Randomized matrix-free quadrature: unified and uniform bounds for stochastic Lanczos quadrature and the kernel polynomial method
• The kernel polynomial method
• The University of Florida sparse matrix collection

This page was built for publication: Accurate Density of States Estimation: A Comparative Study

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7360699)