Innovative MCDM-ML algorithms-based decision-support system for electric vehicle selection

Abstract

Electric vehicles have gained popularity among both manufacturers and consumers in recent years due to government incentives and climate change. In this study, a new decision support system is proposed to optimise the selection of electric vehicles. The proposed system integrates multi-criteria decision making (MCDM) methods with machine learning (ML) algorithms. A dataset consisting of 15 different electric vehicle alternatives and 20 criteria was created. The criteria were weighted using the CRITIC method and the scores of each alternative were determined using the LOPCOW method. The LOPCOW scores were used as target variables for machine learning models and analyses were performed. In the analysis of the machine learning models, performance metrics such as RMSE, MSE, MAE, MAPE and R-2 were calculated and the generalisability of the results was tested using the Kfold5 cross-validation method. Finally, the final ranking of the alternatives was performed using the voting regressor model. The decision support system developed in this study combines CRITIC-LOPCOW machine learning algorithms and voting regressor methods to provide a practical and reliable solution for electric vehicle selection. The findings have significant implications for policymakers, investors, and scholars.