Abstract
We study the regularity of centered Gaussian processes $$(Z_x( \omega ))_{x\in M}$$ , indexed by compact metric spaces $$(M, \rho )$$ . It is shown that the almost everywhere Besov regularity of such a process is (almost) equivalent to the Besov regularity of the covariance $$K(x,y) = {\mathbb E}(Z_x Z_y)$$ under the assumption that (i) there is an underlying Dirichlet structure on M that determines the Besov regularity, and (ii) the operator K with kernel K(x, y) and the underlying operator A of the Dirichlet structure commute. As an application of this result, we establish the Besov regularity of Gaussian processes indexed by compact homogeneous spaces and, in particular, by the sphere.
Highlights
Gaussian processes have been at the heart of probability theory for very long time
There is a huge literature about it. They have been playing a key role in applications for many years and seem to experience an active revival in the recent domains of machine learning as well as in Bayesian nonparametric statistics
Gerard Kerkyacharian LPMA, CNRS-UMR 7599, and Crest E-mail: kerk@math.univ-paris-diderot.fr Shigoyoshi Ogawa Ritsumeikan University, Kyoto:E-mail: ogawa-s@se.ritsumei.ac.jp Pencho Petrushev University of South Carolina E-mail: pencho@math.sc.edu Dominique Picard Universite Paris Diderot - Paris 7, LPMA E-mail: picard@math.univ-paris-diderot.fr. Motivated by these aspects we explore in this paper the regularity of Gaussian processes indexed by compact metric domains verifying some conditions in such a way that regularity conditions can be identified
Summary
Gaussian processes have been at the heart of probability theory for very long time. There is a huge literature about it (see among many others [30], [27], [28] [2], [1] [32]). An important effort has been put on the construction of Gaussian processes on manifolds or more general domains, with the two especially challenging examples of spaces of matrices and spaces of graphs to contribute to the emerging field of signal processing on graphs and extending high-dimensional data analysis to networks and other irregular domains Motivated by these aspects we explore in this paper the regularity of Gaussian processes indexed by compact metric domains verifying some conditions in such a way that regularity conditions can be identified. In this case, the salient fact is not the regularity result (which is known) but the original geometry corresponding to these processes.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have