In this article, we fully characterize the measurable Gaussian processes (U(x))x∈D whose sample paths lie in the Sobolev space of integer order Wm,p(D),m∈N0,1<p<+∞, where D is an arbitrary open set of Rd. The result is phrased in terms of a form of Sobolev regularity of the covariance function on the diagonal. This is then linked to the existence of suitable Mercer or otherwise nuclear decompositions of the integral operators associated to the covariance function and its cross-derivatives. In the Hilbert case p=2, additional links are made w.r.t. the Mercer decompositions of the said integral operators, their trace and the imbedding of the RKHS in Wm,2(D). We provide simple examples and partially recover recent results pertaining to the Sobolev regularity of Gaussian processes.
Read full abstract