Supplemental Material for "Operationalizing a Swiss Cheese Model-Informed Framework for Fatal-Fall Classification: An Interpretable Machine Learning and Web-Based Decision-Support Approach"

This file contains supplementary tables, methodological details, model-validation analyses, and interpretability results supporting the manuscript entitled “Operationalizing a Swiss Cheese Model-Informed Framework for Fatal-Fall Classification: An Interpretable Machine Learning and Web-Based Decision-Support Approach,” submitted to Results in Engineering Journal.The supplementary materials include: • Generalized Variance Inflation Factor (GVIF) analysis for multicollinearity assessment• Hyperparameter optimization settings and computational costs for the evaluated machine-learning models• Statistical model-comparison results using exact McNemar, DeLong, and paired bootstrap analyses• Additional imbalance-sensitive performance metrics, including Balanced Accuracy, MCC, and PR-AUC• Additional SHAP dependence and interaction analyses• Error analysis of the reduced two-feature XGBoost model• Fully nested cross-validation of the reduced-feature XGBoost modeling pipeline• Out-of-time temporal validation using 152 unseen OSHA accident cases from 2024• Computational response-time evaluation of the Streamlit prototype