Machine Learning Modelling of ܥܱଶ Emissions for Sustainable Development: A Comparative Study of Bosnia and Herzegovina, Croatia, an

The growing need for reducing ܥܱଶemissions in the context of sustainable development has intensified the search for efficient analytical approaches to understand and manage emission drivers. In this paper, three machine learning models were developed using multiple linear regression for the countries of Bosnia and Herzegovina, Croatia and Slovenia. Renewable energy consumption, ܲܯଶ,ହ air pollution, ܦܩܲ per capita, foreign direct investment, urban population, forest area, and total population were used as inputs in the models, while ܥܱଶ emissions for the period from 2000 to 2020 were used as outputs. The developed models for all three countries have good performance, with ܴଶvalues of 91,34%, 77,91%, and 77,20% respectively. For Bosnia and Herzegovina urban population increases ܥܱଶemission, while renewable energy consumption and forest area decrease ܥܱଶ emission. In Croatia ܲܯଶ,ହ was the most influential factor that increases ܥܱଶemission.In Slovenia population growth decreases ܥܱଶ emissions, whileGDP per capita increases ܥܱଶ emissions. Also, hypothesis testing for differences between means was performed for all variables between all three countries. The findings showed that for almost all variables there were statistically significant differences in mean differences between all countries. Regarding ܥܱଶ emission there are not enough statistical evidence that Bosnia and Herzegovina have higher ܥܱଶ emissions than Croatia, while both Bosnia and Herzegovina, and Croatia have significantly higher ܥܱଶ emissions than Slovenia. This research shows the potential of machine learning models as tools for data-driven policymaking in the transition towards Industry 5.0 and a sustainable industrial future.