Regularized kernel-based reconstruction in generalized Besov spaces (Q1750389)

Let \(B^{\sigma}_{p,q}(M;D)\) be the generalized Besov spaces which are based on an essentially self-adjoint operator \(D\) and its associated heat kernel defined on a rather general metric measure space \((M, \rho,\mu)\). The embedding properties derived in [\textit{T. Coulhon} et al., J. Fourier Anal. Appl. 18, No. 5, 995--1066 (2012; Zbl 1270.58015)] show that the spaces \(B^{\alpha}_{p,q}(M;D)\) are indeed reproducing kernel Hilbert spaces for \(p=q=2\) and \(\sigma>0\) large enough. An explicit representation of the reproducing kernel is defined by \[ K^{\sigma}(x,y) = \sum_{l=0}^{\infty}b^{-2\sigma l}S^{(l;\sigma)}(x,y), \] where \(b>1\) is a fixed parameter, \(\sigma>0\) determines the smoothness, and \(S^{(l;\sigma)}\) are integral kernels corresponding to operators defined via smooth functional calculus in terms of the operator \(D\). Stability estimates for the kernel, including inverse Bernstein-type estimates for kernel-based trial spaces, are given. Then, explicit coupling relations between the series truncation, the regularization parameters and the data set are derived.