Models of Short-Term Forecasting of Economic Indicators on the Basis of Time Series

Abstract

Under conditions of unstable economy forecasting of economic indicators for timely administrative decision making is of special value. Modern science suggests many methods of forecasting, however not all of them are equally effective in every instance. In the work classification of forecasting methods is presented and their estimation with the reference to current economic situation in Russia is given. Methodical approach to modelling of short-term forecasting of economic indicators on the basis of time series is considered. To illustrate efficiency and universality of the chosen method of short-term forecasting with small amount of statistical data, additive and multiplicative models are constructed with the use of time series of average monthly incomes per head. For considered series factors of autocorrelation of the series levels and seasonal indexes are calculated and analyzed; Modelling with trend methods of regression analysis is made. The analysis of the received forecasting results allowed to make conclusion about high accuracy of considered method and possibility of its use in strategy formation of social and economic development, especially in the situation of economic instability.