Paper Contents
Abstract
The data series tested for stationarity using an augmented dickey-fuller test and data was found non-stationary, and it was analyzed using Autoregressive Integrated Moving Average with explanatory variables model with first differencing, and Feed-Forward Neural Network regression. The ARIMAX model outperformed the feed-forward regression model with leading indicators. The Autoregressive Integrated Moving Average with explanatory variables ARIMAX (1, 1, 1) model predicted the production of Ginger with less root mean square and mean absolute percentage error than the NNAR (1, 2) model, which considers the average of 1000 networks with linear output. The residuals of the fitted models were subjected to the Box-Pierce test and it was found that the residuals from the ARIMAX model were independent while the residuals of NNAR (1, 2) showed dependency over lag. Therefore, the Autoregressive Integrated Moving Average with explanatory variables (ARIMAX) model is the better prediction model over NNAR (1, 2) in terms of performance indicators and residual insignificance.
Copyright
Copyright © 2023 Divyashree . This is an open access article distributed under the Creative Commons Attribution License.
