Novel and refined stability estimates for kernel matrices (2409.04263v1)

Abstract: Kernel matrices are a key quantity within kernel based approximation, and important properties like stability or convergence of algorithms can be analyzed with help of it. In this work we refine a multivariate Ingham-type theorem, which is then leveraged to obtain novel and refined stability estimates on kernel matrices, focussing on the case of finitely smooth kernels. In particular we obtain results that relate the Rayleigh quotients of kernel matrices for kernels of different smoothness to each other. Finally we comment on conclusions for the eigenvectors of these kernel matrices.