Stochastic differential equations arise in a variety of contexts. There are many techniques for approximation of the solutions of these equations that include statistical methods, numerical methods, and other approximations. This article implements a parametric and a nonparametric method to approximate the probability density of the solutions of stochastic differential equation from observations of a discrete dynamic system. To achieve these objectives, mixtures of Dirichlet process and Gaussian mixtures are considered. The methodology uses computational techniques based on Gaussian mixtures filters, nonparametric particle filters and Gaussian particle filters to establish the relationship between the theoretical processes of unobserved and observed states. The approximations obtained by this proposal are attractive because the hierarchical structures used for modeling are flexible, easy to interpret and computationally feasible. The methodology is illustrated by means of two problems: the synthetic Lorenz model and a real model of rainfall. Experiments show that the proposed filters provide satisfactory results, demonstrating that the algorithms work well in the context of processes with nonlinear dynamics which require the joint estimation of states and parameters. The estimated measure of goodness of fit validates the estimation quality of the algorithms.
290APPROXIMATIONS OF THE SOLUTIONS OF A STOCHASTIC DIFFERENTIAL EQUATION nonlinear stochastic differential equations, for example [60] and [5] have used an approach based on a Gaussian process to model the time evolution of the solution of a general stochastic differential equation, the methodology uses data assimilation techniques ([31]). Recently [57] introduced a nonparametric method to estimate the function of drift in a stochastic differential equation where a measure of random probability was considered to model the drift and a Expectation-Maximization (EM) algorithm was developed to try to establish a link between unobserved and observed states. This paper aims to approximate the solutions of a stochastic differential equation by using the Gaussian mixture distribution model, Dirichlet process mixture models, together with a non-parametric density estimation algorithm and three sequential filters. The most representative references about the Gaussian mixture distribution model are [61]; [3]; [68] and [2]. Dirichlet processes were treated in [16], and they have been identified in the literature as a useful tool for estimating densities in the field of non-parametric models. The idea of using Dirichlet process is not new, and has been considered in previous works such as [15]; [46]; [42]; [45]; [70] among others.The rest of article is summarized as follows: Section 2 presents the formulation of the problem; Section 3 contains the description of the filtering algorithms defined to approximate a probability density of the solutions; In Section 4, the results obtained for two different examples are shown, and lastly, Section 5 contains a final discussion an...