Comparison Traditional Methods and Genetic Algorithm in Forecasting Price Fluctuations of Agricultural Selected Products

In current study, the retail price fluctuations of agricultural products including Wheat, Barley, Maize and Rice were predicted by traditional methods and genetic algorithm. Required data for all products from Farvardin 1384 to Esfand 1390 were collected. Aban 1390 to Esfand 1390 data were used for forecasting accuracy and also forecasting was done from Farvardin 1391 to Mehr 1391. In order to compare forecasting error of different methods, root mean square error criterion was used. Results showed that genetic algorithm method among all forecasting methods includes less ererr for forecasting price fluctuations of agricultural selected products, and so it, seasonal autoregressive integrated moving average (SARIMA) method is good. The root mean square error (RMSE) criterion using the genetic algorithm of Wheat, Barley, Maize and Rice price fluctuations are 80.35, 82.78, 376.23 and 923.92 respectively that forecasting price fluctuations of Wheat has the lowest error (80.35). Also, forecasting coming months showed that price fluctuations of selected crops are very high.