Imitation and forecasting of nonstationary financial time series

An approach for modeling of the financial time series with high imitation and forecasting properties has been proposed. Nonstationarity of original time series has been taken into account and the procedure for the reduction of research to the analysis of stationary time series have. The feature of the work is the construction a continuous-discrete models with threedigit variables. The approbation of the developed technique has been performed using real statistical.