Forecasting the price of agriculture products using the decomposition of time series

anglais. Becoming familiar with the price level of particular kind of decisions such as: strategic, tactical, operational should be supported by the results of forecasts with an appropriate time horizon. A short-time method of forecasting agriculture products prices employing a time series decomposition is presented in this study. This method based on the assessment of the mean level of prices resultant from long-period trends and cyclical changes, corrected by seasonal fluctuation and eliminating irregular changes during calculation. The study materials consist of monthly time series of broiler chicken prices for the years 2007-2017. A sample forecast was performed for broiler chicken prices for the time period from one to six month ahead. The level of ex-post relative (MPE) and absolute (MAPE) errors were calculated for each forecast. The presented method can be particularly useful for forecasting prices of agricultural products characterized by seasonal fluctuation.

anglais. price, time series, seasonal fluctuation, broiler chicken

1

165-171

12