首页|Solving Fredholm integral equation of the first kind using Gaussian process regression
Solving Fredholm integral equation of the first kind using Gaussian process regression
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NSTL
Elsevier
Fredholm integral equation of the first kind is a typical ill-posed problem, and it is usually difficult to obtain a stable numerical solution. In this paper, a new method is proposed to solve Fredholm integral equation using Gaussian process regression (GPR). The key to this method is that the right-hand term of the original integral equation is reconstructed by the GPR model to obtain a new integral equation in a reproducing kernel Hilbert spaces (RKHS). We present an analytical approximate solution of the new equation and prove that it converges to the exact minimal-norm solution of the original equation under the L2 -norm. Especially, for the degenerate kernel equation, we obtain an explicit formula of the exact minimal-norm solution. Finally, the proposed method is verified to be very effective in solution accuracy by multiple examples. (c) 2022 Elsevier Inc.
Fredholm integral equation of the first kindDegenerate kernelIll-posed problemGaussian process regressionMoore-Penrose inverseReproducing kernel Hilbert spacesNUMERICAL-SOLUTIONREGULARIZATION
NSTL
Elsevier
