Study of Forecasting Method for Agricultural Products Using Hybrid ANN-GARCH Approach

Abstract

Abstract The capital city of Jakarta as the center of socio-economic activities has to maintain the stability of its economy, including the stability of staple food prices such as rice, onion, and chili prices. To make a strategic plan for anticipatory action, one of the ways is by doing forecasting. The problem in forecasting staple food prices is the fact that those staple foods have high volatility prices. It makes the conditional variance of residual become inconstant. This research applied the Hybrid ANN-GARCH model that combined Artificial Neural Network (ANN) and Generalized Autoregressive Conditional Heteroscedasticity (GARCH) model. This method has used to predict rice, onion, red chili, and cayenne pepper prices in Jakarta. The data divided into two categories, training and testing data. The result shows that the Hybrid ANN-GARCH model produced smaller RMSE and MAPE values than ARIMA model. Based on the accuracy value of the model, it can be concluded that Hybrid ANN-GARCH better than ARIMA in forecasting the price of the four commodities.