Scholarlink

<p>Carbon dioxide (CO<sub>2</sub>) remains the leading contributor to global warming, primarily released through the combustion of fossil fuels in electricity generation, transportation, industry, and residential use. Accurate estimation of CO<sub>2</sub> emissions from different sources is therefore essential for designing effective policies to achieve the Sustainable Development Goals (SDGs). This study aimed to nowcast annual CO<sub>2</sub> emissions in Bangladesh from 1955 to 2021 using various machine learning models, including Extreme Gradient Boosting (XGBoost), Artificial Neural Network (ANN), and Support Vector Model (SVM), and compared them with the traditional ARIMA model. The best-performing model was selected based on root mean square error (RMSE) and mean absolute percentage error (MAPE). Among the models, SVM achieved the best performance for oil, cement, and gas, while ANN outperformed others for coal. These findings demonstrate that machine learning models provide the most accurate nowcasting compared to ARIMA, and it is recommended to consider a machine learning model for the nowcasting of CO<sub>2</sub> emissions in Bangladesh.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Nowcasting CO2 emissions in Bangladesh: a machine learning approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月13日

查看全文
  • Md. Moyazzem Hossain,
  • Faruq Abdulla,
  • Azizur Rahman,
  • Md Galal Uddin,
  • Agnieszka I. Olbert

摘要

Carbon dioxide (CO2) remains the leading contributor to global warming, primarily released through the combustion of fossil fuels in electricity generation, transportation, industry, and residential use. Accurate estimation of CO2 emissions from different sources is therefore essential for designing effective policies to achieve the Sustainable Development Goals (SDGs). This study aimed to nowcast annual CO2 emissions in Bangladesh from 1955 to 2021 using various machine learning models, including Extreme Gradient Boosting (XGBoost), Artificial Neural Network (ANN), and Support Vector Model (SVM), and compared them with the traditional ARIMA model. The best-performing model was selected based on root mean square error (RMSE) and mean absolute percentage error (MAPE). Among the models, SVM achieved the best performance for oil, cement, and gas, while ANN outperformed others for coal. These findings demonstrate that machine learning models provide the most accurate nowcasting compared to ARIMA, and it is recommended to consider a machine learning model for the nowcasting of CO2 emissions in Bangladesh.