Zohdy, MohamedAlbariqi, IbrahimSiadat, Mohammad-RezaEdwards, WilliamSmith, Julia2026-09-282026-09-282026-01-01https://hdl.handle.net/10323/22206Global supply chains have become increasingly interconnected and data-intensive, but recent disruptions have shown that traditional supplier selection approaches, often focused on cost, quality, and delivery, are insufficient for ensuring continuity and responsible sourcing. Small and medium-sized enterprises (SMEs) are especially vulnerable because they typically lack the analytical capacity to evaluate suppliers using both resilience and sustainability criteria. This study develops andevaluates a machine learning based decision support system, the Resilient-Sustainable Supplier Selection System (RSSSS), to support data-driven supplier classification and risk-aware sourcing decisions. The proposed RSSSS integrates operational performance indicators with resilience attributes (e.g., reliability, agility, and geographic diversification) and sustainability-related indicators within a unified evaluation framework. A historical supplier dataset containing 74,339 records and 13 predictive features was preprocessed through data cleaning, categorical encoding, and handling of missing values. Seven supervised machine learning models—Decision Tree, Random Forest, AdaBoost, Naive Bayes, k-Nearest Neighbors, Linear Discriminant Analysis, and XGBoost—were trained and evaluated using accuracy and computational training time. Results show that XGBoost achieved the highest classification accuracy (82.23), outperforming all alternative models, while requiring 5.96 seconds of training time. The findings confirm that tree-based ensemble learning can effectively capture complex, non-linear relationships among resilience and sustainability indicators, enabling more accurate identification of resilient-sustainable suppliers than traditional rule-based or static scoring methods. Overall, the RSSSS demonstrates a practical and lightweight approach for SMEs to improve supplier selection speed, transparency, and comprehensiveness, supporting more resilient, sustainable, and risk-aware sourcing decisions under uncertainty.