Search

Minimizing the detrimental effects of road transport greenhouse gas (GHG) emissions on climate change and global warming requires accurate emission forecasting. To forecast greenhouse gas emissions from industrial and civilian transportation on roads in China, we present new approaches that use data extraction and managed machine learning methods for regression and identification. Four methods are examined: decision tree, multinomial logistic regression, multivariate linear regression, and artificial neural network (ANN) multiple-layer perceptron. The findings suggest that the multiple-layer perceptron approach of ANN has superior prediction accuracy compared to other models. Ensemble modelling techniques, such as Bagging and Boosting, significantly improved the predictive performance of the developed multilayer perceptron system. The paper's conclusions are significant for transport policymakers, regulators, and international organizations in mitigating GHG emissions. [ABSTRACT FROM AUTHOR]

Exposure to airborne pollution has substantial adverse health consequences. Governments around the world have paid attention to this problem and started to take action to mitigate its harmful effects. In this paper, we investigate how a change in the consumption tax structure affects car emissions by exploiting exogenous variation from a natural experiment that took place in China. Our results show that this tax policy, which doubled the imposition paid on cars with large engines, reduced the emissions of all the pollutants studied, with the most significant decrease of 11% noted in Particulate Matter and Carbon Monoxide. These results have strong policy implications. By showing the effectiveness of this policy, we uncover the possible benefits of replicating it in different contexts. Further, by using information from COPERT, a totally new approach to evaluate tax and environmental policies, we propose new tools for policymakers to generate rich and detailed data that can be used in different contexts to evaluate policies where data from the population of automobiles is needed. [ABSTRACT FROM AUTHOR]