The study examines the theoretical and practical issues of forecasting financial and economic performance of enterprises with a seasonal nature of production. The paper determines the distinctive features of the basic econometric prediction models – additive and multiplicative, based on the nature of seasonal fluctuations. The authors have analyzed and compared the applied models designed to forecast seasonal processes – those include numerous adaptive models, a seasonal version of the auto-regression model of an integrated moving average, and trend-seasonal models. The study examines the advantages and disadvantages of trend-seasonal models compared to other econometric forecasting models addressing its seasonal component. The authors suggest a modified algorithm for constructing trend-seasonal models, which, on the one hand, allows increasing the accuracy of a typical algorithm, and, on the other hand, considers the functionality of modern tabular processors for constructing trend-seasonal models by nonprogrammer users. The practical testing of the proposed algorithm was implemented by predicting the profit of an enterprise with a seasonal nature of production. The authors also conducted a comparative evaluation of the accuracy of the proposed algorithm with a typical algorithm for constructing trend-seasonal models.
Volume 11 | 08-Special Issue
Pages: 935-947