Адаптивные стохастические модели в трудно формализуемых естественнонаучных задачах

Abstract

To formalize and estimate the dynamical processes in the difficultly formalized problems of the different physical nature, an approach to create the stochastic models with the a priori unknown noise statistics based on the approximating determinate functions is proposed. The examples of the development of models for the estimation and prediction of the solar activity and the blood pressure based on the identification of the noise statistics and the adaptive filtration are given.