Application of garch and ann models for Potato Price Forecasting: A case study of Bangalore market, Karnataka state Areef M, Radha Y Department of Agricultural Economics, Acharya N. G. Ranga Agricultural University, Lam, Guntur-522034. JEL Codes: C01, C45, C53 and C87 Online published on 27 May, 2021. Abstract The present study is an attempt to modeling and forecasting the prices of potato using Generalized Auto Regressive Conditional Heteroscedasticity (GARCH) model and Artificial Neural Network (ANN) model at Bangalore market in Karnataka state. The GARCH and ANN models have been compared in terms of lower values of mean average percentage error (MAPE), root mean square error (RMSE), mean absolute error (MAE) and mean absolute scaled error (MASE). The monthly prices of potato from January, 2005 to December, 2019 were used to train the model and data pertaining to the period January, 2020 to August, 2020 were used as test data to finalize the model for forecasting. The ANN model showed that forecasted prices of potato were very close to actual prices as compared to GARCh model at Bangalore market in Karnataka state. ANN model predicted high future prices for the month of January, 2021 ( 2247 per quintal) and lower prices for the month of September, 2020 ( 1024 per quintal). Predicted future prices help the farmers in adjusting the sowing and harvesting timings in order to ensure better price for potato in the market. Forecasted prices are useful to governments to regulate exports and storage quotas of traders to make timely availability of potato to consumers at fair prices. Top Keywords ANN, GARCH, MAPE, MASE and RMSE. Top |