Forecasting carbon dioxide emissions in Turkey using machine learning methods

Global warming and climate change are among the most important problems that will affect our future and threaten it seriously. Carbon dioxide gas is emitted into the atmosphere using fossil fuels. Therefore, international organisations have accelerated policies that reduce carbon emissions to mitigate the effects of global warming. In this study, CO2 emissions in Turkey were estimated by machine learning methods. A total of 34 data were used for analysis between 1980 and 2014. Annual average temperature, population, gross domestic product (GDP), industry (annual % growth), electricity consumption (kWh per person), coal consumption (thousand tons), amount of agricultural land (km(2)), oil consumption (barrel per day) information was used in Turkey. The order of success of the machine learning methods used in the study was artificial neural networks, support vector machines and decision trees.