Traffic Conflict Identification and Risk Assessment at Ramp Merging Areas Based on Improved XGBoost and SHAP

Abstract

AbstractRamp merging areas are high-risk zones for traffic accidents due to the high traffic density and frequent changes in vehicle speed. This study proposes a comprehensive approach that integrates machine learning with feature interpretation to ...Practical ApplicationsRamp merging areas, where vehicles enter or exit highways, are known to be high-risk zones for accidents due to heavy traffic and frequent speed fluctuations. This study offers a new method for assessing the risk of traffic conflicts ...