Abstract

Carbon dioxide (CO2) emissions is among of global environmental pollutants contributing to climate change. The current study aims to create an Autoregressive Integrated Moving Average with external factors (ARIMAX) model to predict CO2 emissions in Tanzania. In this study, an Autoregressive Integrated Moving Average (ARIMA) model is first created. Then it is combined with the influencing factors using multiple linear regression to fit an ARIMAX model. There is a high possibility of both under- and overestimation because the ARIMAX employing multiple linear regression (ARIMA-MLR) model only generates mean forecasts. A hybrid ARIMA-Quantile Regression (ARIMA-QR) is created to forecast high and low quantiles. The ARIMA-QR model mainly predicts the quantiles instead of extrapolating from the mean point of the ARIMA-MLR model, which bases more on the assumption of normality. The established ARIMA-MLR and ARIMA-QR were used to forecast and model annual data on CO2 emissions in Tanzania. The findings reveal that both ARIMA-MLR and ARIMA-QR models outperform the traditional ARIMA model in terms of forecasting accuracy with the least mean absolute percentage error (MAPE) and root mean square error (RMSE).