Enhanced Gaussian processes and applications
نویسندگان
چکیده
We propose some construction of enhanced Gaussian processes using Karhunen-Loeve expansion. We obtain a characterization and some criterion of existence and uniqueness. Using roughpath theory, we derive some Wong-Zakai Theorem. Mathematics Subject Classification. 60G15, 60G17. Received July 2, 2007. Revised October 5, 2007. 1. Generalities In [13] Lyons developed a general theory of differential equations of the form dyt = f(yt)dxt. (1.1) Classical integration/ODE theory gives a meaning to such differential equations when x has bounded variation. Lyons extended this notion to the case when x is a path with values in a Banach space B, and of finite p-variation, p ≥ 1. To do so, one needs first to lift x to a path of finite p-variation in the free nilpotent group of B. In other words, one needs to define and make a choice for the “iterated integrals” of order less than or equal to [p] of x. We refer the reader to, for example, [11,13,14]. In this paper, our aim is to work towards the study of a “natural” p-rough path process lying above an arbitrary Gaussian process. We simplify the problem by only looking at lift in the free nilpotent group of step 2, i.e. we are just looking at the Lévy area of Gaussian processes. This was already done by Lévy in 1950 for Brownian motion, see [12] or more recently [10] and [6], and for fractional Brownian motion, see [5]or [15]. Moreover, Biane and Yor, in [1] have constructed the Lévy area using the expansion of Brownian motion in the basis of Legendre polynoms. Karhunen-Loeve expansion Theorem provide a natural way to approximate paths of a Gaussian process by a smooth process. This paper is devoted to study how its expansion allow to lift R-valued Gaussian process x to a path x with values in some free nilpotent of step 2 group over R (or in other words, how to construct the Lévy area of x, i.e. the second iterated integral of x). We also show that if the process x with some area process satisfies some quite natural conditions, then x will be the limit of the lift of the Karhunen-Loeve approximations of x. The proof of the convergence of Karhunen-Loeve expansion Theorem or of some properties on Gaussian processes relies on the convex property of the vector spaces. The free nilpotent group of step 2 do not share this property. In the first part of this paper, we give a proof of a weak version the Karhunen-Loeve expansion
منابع مشابه
The Rate of Entropy for Gaussian Processes
In this paper, we show that in order to obtain the Tsallis entropy rate for stochastic processes, we can use the limit of conditional entropy, as it was done for the case of Shannon and Renyi entropy rates. Using that we can obtain Tsallis entropy rate for stationary Gaussian processes. Finally, we derive the relation between Renyi, Shannon and Tsallis entropy rates for stationary Gaussian proc...
متن کاملComplete convergence of moving-average processes under negative dependence sub-Gaussian assumptions
The complete convergence is investigated for moving-average processes of doubly infinite sequence of negative dependence sub-gaussian random variables with zero means, finite variances and absolutely summable coefficients. As a corollary, the rate of complete convergence is obtained under some suitable conditions on the coefficients.
متن کاملOn the Maximum Likelihood Training of Gradient-Enhanced Spatial Gaussian Processes
Abstract. Spatial Gaussian Processes, alias spatial linear models or Kriging estimators, are a powerful and wellestablished tool for the design and analysis of computer experiments in a multitude of engineering applications. A key challenge in constructing spatial Gaussian processes is the training of the predictor by numerically optimizing its associated maximum likelihood function depending o...
متن کاملADK Entropy and ADK Entropy Rate in Irreducible- Aperiodic Markov Chain and Gaussian Processes
In this paper, the two parameter ADK entropy, as a generalized of Re'nyi entropy, is considered and some properties of it, are investigated. We will see that the ADK entropy for continuous random variables is invariant under a location and is not invariant under a scale transformation of the random variable. Furthermore, the joint ADK entropy, conditional ADK entropy, and chain rule of this ent...
متن کاملA Review of Electrospun Nanofber Membranes
Electrospun nanofber membranes (ENMs) are new generation of membranes with many favorable properties such as high flux and low pressure drop. Although electrospinning has been known for more than a century, its applications in fltration and separation processes are relatively new. Electrospinning has provided the means to produce ultrathin fbers – as thin as a few nanome...
متن کاملStudent-t Processes as Alternatives to Gaussian Processes
We investigate the Student-t process as an alternative to the Gaussian process as a nonparametric prior over functions. We derive closed form expressions for the marginal likelihood and predictive distribution of a Student-t process, by integrating away an inverse Wishart process prior over the covariance kernel of a Gaussian process model. We show surprising equivalences between different hier...
متن کامل