Development of model of prognostication of temporal rows with heteroskedastic and volatility with the use of neuron networks

摘要

the problem of choice of method of prognostication taking into account statistical descriptions of basic data is considered In work. The mathematical model of prognostication of temporal rows with heteroskedastic in the conditions of volatility basic data is developed. Statistical descriptions which the choice of method of prognostication in the conditions of heteroskedastic is possible on the basis of are determined, a neuron network for the search of optimum method of prognostication, which was trained through the data got on the basis of imitation design, is Offered and built.