<![CDATA[The escalation of carbon dioxide (CO2) emissions has emerged as a critical environmental concern, particularly in the context of Indonesia’s pursuit of sustainable development. This study aims to forecast CO2 emissions in Indonesia using annual time-series data spanning 1967–2023. Three methodological approaches are employed: an artificial neural network (ANN), an autoregressive model with exogenous variables (ARIMAX), and a hybrid ARIMAX-ANN model. The dataset comprises Gross Domestic Product obtained from the World Bank, along with per capita CO2 emissions, per capita natural gas consumption, and per capita hydropower consumption sourced from Our World in Data. The findings of this research demonstrate that the hybrid ARIMAX-ANN model provides the best forecasting performance, as evidenced by the lowest RMSE, MAPE, and MAE values among the other two models. These results suggest that the hybrid model is currently the most reliable for predicting CO2 emissions in the Indonesian context. The study enriches the expanding literature on emission forecasting by providing empirical evidence to support data-driven policymaking for climate change mitigation and sustainable energy development in Indonesia. ]]>
Copyrights © 2025
