A Comparative Analysis of Machine Learning Model Utilization for the Optimization of Supplier Reliability Towards Sustainable Construction

Engineering, Technology & Applied Science Research (ETASR) 2025
Jānis Zvirgzdiņš, Thilina Ganganath WEERAKOON, Ineta Geipele, Keshmin Senaratna

The success of a construction project is significantly influenced by the reliability of its material suppliers, particularly when operational efficiency and sustainability factors are considered. Traditionally, project managers have relied on conventional scorecards and heuristic decision-making methods, which are often subjective and limited in scope. To provide a more objective and data-driven approach, this study introduces a Machine Learning (ML)-based pipeline for evaluating supplier reliability. Four ML algorithms, Random Forest (RF), Extreme Gradient Boosting (XGBoost), Support Vector Machine (SVM), and K-Nearest Neighbors (KNN), were assessed, while Shapley Additive Explanations (SHAP) were employed to interpret model outputs and identify the supplier characteristics most influential to each algorithm. Additionally, a proprietary Sustainability Score (SC) was integrated to account for the sustainability aspect provided by each supplier. The dataset used consisted of 105 records from five major suppliers involved in two large-scale residential construction projects in Sri Lanka, incorporating key features such as waste generation, lead time, material cost, and delivery accuracy. Among the evaluated models, XGBoost demonstrated the best performance, achieving an F1-score of 0.89 and an Area Under the Curve (AUC) of 0.97, followed by RF. Notably, the SC analysis revealed that some medium-reliability suppliers achieved higher sustainability scores than high-reliability ones, highlighting the importance of multi-criteria evaluation frameworks that balance operational reliability with environmental responsibility in supplier selection. By facilitating precise, open, and ecologically responsible supplier evaluations, this integrated ML framework aids in sustainable procurement decisions for construction projects.

Keywords
artificial intelligence, machine learning, supply chain management, sustainable construction
DOI
10.48084/etasr.12724

Zvirgzdiņš, J., Weerakoon , T.G., Geipele, I., Senaratna, K. A Comparative Analysis of Machine Learning Model Utilization for the Optimization of Supplier Reliability Towards Sustainable Construction. Engineering, Technology & Applied Science Research (ETASR), 2025, Vol. 15, No. 6, pp. 29907-29913. ISSN 2241-4487. e-ISSN 1792-8036. Available from: doi:10.48084/etasr.12724

Publication language
English (en)