Application of Decision Tree and Machine Learning in New Energy Vehicle Maintenance Decision Making

Abstract

Abstract Several incidents of electric vehicle combustion across various regions in China have brought the safety concerns associated with new energy vehicles into sharp focus within public discourse. Addressing these concerns, this paper explores maintenance decision-making for new energy vehicles through the application of decision trees and machine learning techniques. Initially, the study analyzes how decision trees and machine learning are employed in crafting maintenance decisions for these vehicles. It involves collecting data through internet searches, followed by statistical analyses and preprocessing to set the groundwork for further inquiry. Furthermore, the research advances by developing and refining decision tree models, which facilitate the integration of fault diagnosis and maintenance decision-making processes for new energy vehicles. This effort culminates in the establishment of a robust decision tree model specifically designed for the maintenance of new energy vehicles, which is subsequently evaluated through a detailed case study. The results are presented: the approximation degree of new energy vehicle fault diagnosis based on the decision tree model is 90.58%, 90.67%, 88.09%, 91.28%, and 90.19% at the significant level of α taking the value of 0.01, 0.05, 0.100, respectively. This study provides theoretical guidance on the application of decision trees for the diagnosis and repair of faults in new energy vehicles so that the development of new energy vehicles in China can be ensured.