Fig 1.
Theoretical framework for the study.
Fig 2.
Boxplot of the annual rice production data in Bangladesh from 1961 to 2020.
Fig 3.
A time series plot for rice production in Bangladesh from 1961 to 2020.
Fig 4.
A comparison between the Box-Cox transformed sequence and the original sequence of annual rice production in Bangladesh.
Fig 5.
First-order differencing of the rice production of the training data set shows stationarity.
Fig 6.
The ACF and PACF diagram of rice production in Bangladesh after first order differencing.
ACF, autocorrelation function; PACF, partial autocorrelation function.
Table 1.
Estimated parameters of the ARIMA (0,1,1) with drift model.
Fig 7.
A time series plot of the residuals with corresponding ACF diagram, and a histogram for the ARIMA (0,1,1) model with drift.
ACF, autocorrelation function; ARIMA, autoregressive integrated moving average.
Fig 8.
Important characteristic features of the XGBoost model.
Fig 9.
ARIMA and XGBoost model show the actual, fitted and forecasted data for rice production in Bangladesh.
ARIMA, autoregressive integrated moving average; XGBoost, eXtreme Gradient Boosting.
Table 2.
Evaluation of parameters for the ARIMA and XGBoost model for rice production in Bangladesh.
Fig 10.
Ten years’ prediction of annual rice production in Bangladesh using XGBoost model.
XGBoost: eXtreme Gradient Boosting.