РОЗРОБКА МОДЕЛІ ДЛЯ ВИРІШЕННЯ ЗАДАЧІ ПРОГНОЗУВАННЯ

Parametric and non-parametric models suitable for the description of the processes, which are stationary in broad sense, are considered. The application of neural network models is shown to be quite effective, particularly, the parallel and series-parallel models based on either the multilayer perceptrons or radial basis or generalized regression networks