Paper Contents
Abstract
In the present study, an attempt was made to forecast the Ginger Area and Production using Artificial Neural Networks and Box-Jenkins ARIMA models, and compared both the models on the basis of accuracy criterias. The ANN models utilize back propagation and conjugate gradient methods to train the data in the hidden layers, and sigmoid function as a transformation function. The production of ginger was predicted closely by Neural Network 1:2-20:1 with less training (0.168) and testing error (0.051) with R-Square value 0.82 and ARIMA (2, 0, 3) forecasted the production with R-Square value 0.77. For predicting the area of ginger, ANN 1:2-7:1 outperformed the ARIMA (3, 0, 4) models with low RMSE value, lesser training (0.027) and testing errors (0.452). The ANN models were better in learning the complexity of the data series and to predict the out sample forecasts. Therefore, ANN model could be used as forecasting technique to get the time series projection of production and area of ginger for India.
Copyright
Copyright © 2023 Divyashree. This is an open access article distributed under the Creative Commons Attribution License.
