Comparison of Artificial Intelligence Methods: Decision Tree vs. Support Vector Machine for Assessing Supply Chain Resilience in Automotive Parts

Abstract

This study aims to enhance the resilience of the automotive parts supply chain and compare the effectiveness of artificial intelligence techniques, specifically Decision Tree and Support Vector Machine (SVM) models. The research dataset consists of 200 simulated records from various supply chain scenarios. For each sample, indicators such as the number of suppliers, average delivery time, safety stock, disruption frequency, and response speed were measured. Model performance was evaluated based on metrics including Accuracy, Recall, F1-Score, and the Confusion Matrix. The results revealed that the Decision Tree model, with an accuracy of 0.92, recall of 0.91, and F1-score of 0.92, demonstrated superior classification capability compared to SVM. While SVM achieved close performance with an accuracy of 0.91 and recall of 0.90, it was less effective in terms of interpretability and decision-making transparency. Additionally, in terms of AUC in the ROC curve and the Precision–Recall metric, the Decision Tree model outperformed the SVM. Beyond its higher accuracy, the Decision Tree model offered greater advantages in identifying influential factors affecting supply chain resilience and in providing transparent decision-making pathways. In contrast, SVM proved more effective in analyzing complex patterns and nonlinear data, although it suffered from lower interpretability. Overall, the findings of this study confirm that artificial intelligence techniques contribute to improved resilience, risk management, and decision optimization in the automotive parts supply chain. Based on the results, it is recommended to implement policies such as supplier diversification, intelligent safety stock management, and enhancement of disruption response speed to bolster the supply chain's robustness against diverse challenges.