Track: Machine Learning

Abstract

This study generally aimed to develop a model that can predict carbon footprint emissions based on the datasets obtained. The parameters that the researchers utilized were population and energy production from 1990 to 2019. The data were all chosen from the vital drivers of CO2 emissions identified through the literature review and had been collected using the open-sourced program WEKA. The three algorithms used in this study are Linear Regression (LR), Multilayer Perceptron (MLP), and Sequential Minimal Optimization Regression (SMOreg) to test which among these most-used algorithms in time series data mining prediction delivers outstanding results. These were participated in the trials given to the researchers to gather sufficient data to draw conclusions. The researchers determined that the MLP model is able to estimate greenhouse gas emissions at error levels that are acceptable in comparison to the actual values and test results. If the Philippines maintains its current trajectory, the carbon footprint emissions will eventually rise and probably reach 255,595.03 kt in the year 2030. There is a powerful correlation between population and energy production that contributes to the carbon footprint emissions within the Philippines.