Abstract:In this paper we investigate the conditional distributions of two Banach space valued, jointly Gaussian random variables. We show that these conditional distributions are again Gaussian and that their means and covariances are determined by a general finite dimensional approximation scheme based upon a martingale approach. In particular, it turns out that the covariance operators occurring in this scheme converge with respect to the nuclear norm and that the conditional probabilities converge weakly. Moreover, we discuss in detail, how our approximation scheme can be implemented in several classes of important Banach spaces such as (reproducing kernel) Hilbert spaces and spaces of continuous functions. As an example, we then apply our general results to the case of Gaussian processes with continuous paths conditioned to partial but infinite observations of their paths. Here we show that conditioning on sufficiently rich, increasing sets of finitely many observations leads to consistent approximations, that is, both the mean and covariance functions converge uniformly and the conditional probabilities converge weakly. Moreover, we discuss how these results improve our understanding of the popular Gaussian processes for machine learning.

Conditional Expectations in Banach spaces with RNP

Marcinkiewicz's Strong Law of Large Numbers for Nonlinear Expectations

Conditional expectations of random holomorphic fields on Riemann surfaces

Conditioning of Banach Space Valued Gaussian Random Variables: An Approximation Approach Based on Martingales

Conditional supremum in Riesz spaces and applications

Strong Limit Theorems for Extended Independent Random Variables and Extended Negatively Dependent Random Variables under Sub-Linear Expectations

The Sufficient and Necessary Conditions of the Strong Law of Large Numbers under Sub-Linear Expectations

Probability Bracket Notation: Probability Space, Conditional Expectation and Introductory Martingales

Rosenthal’s Inequalities for Independent and Negatively Dependent Random Variables under Sub-Linear Expectations with Applications

Lindeberg's Central Limit Theorems for Martingale-Like Sequences under Sub-Linear Expectations

Sublinear Expectations and Martingales in discrete time

Moment and exponential estimation for the distribution of the norms for random matrices martingales

Splitting of conditional expectations and liftings in product spaces

Integral representation of martingales motivated by the problem of endogenous completeness in financial economics

Generating functions in Riesz spaces

Conditional probability in Renyi spaces

Some Estimates for Martingale Representation under G-Expectation

A categorical treatment of the Radon-Nikodym theorem and martingales

Optimal Estimates for Martingale Transforms

Approximations of expectations under infinite product measures

Sequences in non-commutative L^p-spaces