A COMPARISON OF THE USEFULNESS OF WINTERS’ AND SARIMA MODELS IN FORECASTING OF PROCUREMENT PRICES OF MILK IN POLAND

In the paper the Winters’ model has been studied as one from adaptive models based on exponential smoothing methods as well as seasonal autoregressive integrated moving average model SARIMA. The aim of the paper is the assessment of accuracy of short-term forecasts of procurement prices of milk in Poland. Empirical verification of ex post forecasts of monthly procurement prices of milk on the basis of 109 time series with 12-month forecast horizon was conducted. Forecasts constructed with the use of SARIMA model are more often exact than when additive and multiplicative Winters’ model are used.