EXPLORING THE INTERACTION EFFECTS OF KEY INDICATORS FOR CARBON EMISSION THROUGH PENALIZED REGRESSION MODELS

Abstract

Carbon emissions, primarily in the form of carbon dioxide (CO₂), play a critical role in climate change and its associated environmental consequences. This study explores the causes, effects, and potential solutions related to carbon emissions, highlighting their significance in global efforts to mitigate climate change. Pakistan's carbon emissions primarily originate from energy production, industrial activities, transportation, and agricultural practices. The current study focused on the efficient model selection for Carbon emission in Pakistan. The CO₂ (Solid, Liquid, Gaseous) fuel consumption is taken as the response variable while the Population, Total Value, Agriculture, forestry, and fishing, value added (% of GDP) Value, Agricultural land (sq. km) Value, Urban population Value, Gross fixed capital formation (current US$), Industry (including construction), Fertilizer consumption (% of fertilizer production) and GDP taken as the predictors. Five stages are performed for the efficient model selection. In the first stage, multicollinearity diagnosis is applied in term of correlation matrix. For the second stage, outlier diagnosis is performed through the different statistical measures with the graphical analysis in term of Box Plot. The third stage was based on the best model selection by using Ridge and LASSO regression analysis. The fourth stage was consisted on the efficient model selection by using the model selection criteria as Sum Square of Error (SSE), Mean Square Error (MSE), Root Mean Square Error (RMSE) and forecasting efficiency is tested through the Mean Absolute Percentage Error (MAPE). The study result shows that Ridge regression analysis is selected as the best technique on the basis of maximum  , minimum MSE, MAPE and LASSO regression.