English

On the speed of uniform convergence in Mercer's theorem

Machine Learning 2022-09-27 v2 Spectral Theory

Abstract

The classical Mercer's theorem claims that a continuous positive definite kernel K(x,y)K({\mathbf x}, {\mathbf y}) on a compact set can be represented as i=1λiϕi(x)ϕi(y)\sum_{i=1}^\infty \lambda_i\phi_i({\mathbf x})\phi_i({\mathbf y}) where {(λi,ϕi)}\{(\lambda_i,\phi_i)\} are eigenvalue-eigenvector pairs of the corresponding integral operator. This infinite representation is known to converge uniformly to the kernel KK. We estimate the speed of this convergence in terms of the decay rate of eigenvalues and demonstrate that for 2m2m times differentiable kernels the first NN terms of the series approximate KK as O((i=N+1λi)mm+n)\mathcal{O}\big((\sum_{i=N+1}^\infty\lambda_i)^{\frac{m}{m+n}}\big) or O((i=N+1λi2)m2m+n)\mathcal{O}\big((\sum_{i=N+1}^\infty\lambda^2_i)^{\frac{m}{2m+n}}\big). Finally, we demonstrate some applications of our results to a spectral charaterization of integral operators with continuous roots and other powers.

Keywords

Cite

@article{arxiv.2205.00487,
  title  = {On the speed of uniform convergence in Mercer's theorem},
  author = {Rustem Takhanov},
  journal= {arXiv preprint arXiv:2205.00487},
  year   = {2022}
}