What drives perceived safety concerns at roundabouts under disordered, heterogeneous traffic? An inquiry using explainable machine learning

Evaluating roundabout safety in low- and middle-income countries is challenging due to unreliable crash data. This study addresses the issue by analyzing safety perception data from 1530 questionnaire respondents across two Indian cities. A novel framework was developed to compare advanced ordinal ensemble models (Ordered XGBoost, LightGBM, and Random Forest) against conventional ordinal regression (Logit, Probit). The ensemble models proved vastly superior, achieving Quadratic Weighted Kappa (QWK) scores exceeding 0.94, while conventional methods scored below 0.53. The top-performing Ordered XGBoost model (QWK = 0.97) was interpreted using the state-of-the-art explainable artificial intelligence technique SHAP (SHapley Additive exPlanations). SHAP analysis quantified the influence of key factors on perceived risk, identifying personal attributes (occupation, accident/near-accident experience), and infrastructure deficits (inadequate lighting, missing navigational aids) as primary drivers. The findings offer SHAP-quantified insights for deploying targeted, evidence-based safety interventions, providing a blueprint for improving perceived safety in complex traffic environments where traditional analysis is infeasible.