Application of exponential smoothing models in forecasting high frequency time series in the condition of lack of full information

The paper will present the results of the application of the modified additive and multiplicative exponential smoothing models (Brown, Holt and Holt-Winters) in the interpolation and extrapolation forecasting of demand for power energy in the agglomeration A in hour periods, based on time series with systematic gaps. The basis for the construction of forecasts will be time series, from which twelve month, weekly and twenty-four hour fluctuation cycles have been eliminated. Additionally the comparative analysis of accuracy of forecasts built for classical time series models with complex seasonal fluctuations will be conducted. There also will be presented an assess of the criteria for selecting the optimal values of the smoothing constants in terms of building an ex ante forecasts