A hybrid intelligent predicting model for exploring household CO2 emissions mitigation strategies derived from butterfly optimization algorithm

Abstract

The warming on earth is having a profound impact on human survival. With the improvement of people’ living standard, the consumption of energy in residential sector has raised swiftly, leading to a rapid increase in corresponding CO2 emissions. To effectively mitigate household emissions, taking the Yangtze River Delta (YRD) region in China as a case study, this paper proposes a novel intelligent model combining driving forces exploration and prediction. The work first estimates the residential energy-related CO2 emissions precisely, and then the bivariate correlation analysis will be applied to analyze region discrepancy in main affecting factors of emissions based on 13 preliminary indicators. To obtain the principal information of above influencing factors as the input of prediction model, the kernel principal component analysis (KPCA) is introduced innovatively. Besides, butterfly optimization algorithm (BOA) is enhanced to better optimize the parameters of least square support vector machine (LSSVM). The new proposed hybrid model, hereafter called as EBOA-LSSVM, will be utilized to predict residential CO2 emissions in the YRD. Ultimate simulation results present the new model's prominent performance through comparing prediction accuracy with other models. Finally, this article provides some advice for policy makers to guide CO2 emissions reduction from residents department.