Prediction of energy's environmental impact using a three-variable time series model

From the Vienna Convention for the Protection of the Ozone Layer (VCPOL) held in 1985 to the joint United Nations Convention of Climate Change held in 2010, environmental protection has become increasingly urgent. In 2011, the International Standards Organization launched a new energy management system ISO50001 for improving energy efficiency. As an adjunct to these increasingly stringent international agendas, a three-variable time series model is proposed to improve the prediction accuracy of related data. Using a 29 year (1982 to 2010) panel data in Taiwan, the model shows that the environmental impact of increases in CO₂ emission is highly correlated to the three leading impact factors, i.e. GDP per capita, renewable energy supplies and coal consumption. The proposed method integrates these three leading impact factors into a highly accurate stationary prediction model. Then, the future environmental impact is evaluated by preforming the trend analysis of CO ₂ emissions with different growth rate combinations of coal consumption, renewable energy supplies and GDP per capita. Finally, a comparative analysis is performed between our proposed method and the backpropagation neural network (BPN). The comparison results show that the prediction accuracy of our proposed method outperforms BPN in terms of mean absolute percentage error (MAPE) and mean absolute scaled error (MASE).