Learning Closure of Dynamical Systems with Kernel Ridge Regression

DOI10.48550/arXiv.2610.02564MaRDI QIDQ7367479arXivFDO

Authors Evan Habbershaw, John Harlim, Senwei Liang

Publication date 1 October 2026

Published in arXiv

Tldr We propose a closure modeling framework using Kernel Ridge Regression (KRR) to identify missing components of dynamical systems. Our contribution is a novel KRR-based approach that accurately predicts long-horizon solutions for difference-equation closures in ODE and PDE settings, with error bounds quantifying the impact of time integration, approximation, and interpolation.

Full work available at URL https://doi.org/10.48550/arXiv.2610.02564

arXiv classification

cs.LGmath.DSmath.NA

This page was built for publication: Learning Closure of Dynamical Systems with Kernel Ridge Regression

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7367479)