Forecasting CO2 Emissions from Energy, Manufacturing and Transport Sectors in Pakistan: Statistical Vs. Machine Learning Methods

Abstract

The literature of energy forecast is moving from the traditional statistical prediction methods to today’s more intelligent machine learning algorithms due to continuous development in these prediction methods. But, in the literature of forecasting, the superiority of the machine learning algorithms is scarce concerning their relative performance compared to traditional statistical forecasting methods. The purpose of this study is to evaluate the performance of statistical and machine learning methods for forecasting CO2 emissions from energy, transport and manufacturing sectors of Pakistan and provide sector wise CO2 mitigation strategies. We projected the CO2 emissions sector wise using selected methods till 2030 and also gave suggestions for the policymakers to make better decisions. The results show the increasing trend in CO2 emission from energy, manufacturing and transport sectors in Pakistan which causes the significant problems if the government fail to respond sector related issues appropriately.