Modeling dispositional and initial learned trust in automated vehicles with predictability and explainability

• XGBoost can predict dispositional and initial learned trust in automated vehicles. SHAP explain prediction model of both locally globally. Benefits risks are critical to Technological advances the automotive industry bringing driving closer road use. However, one most important factors affecting public acceptance vehicles (AVs) is public’s AVs. Many influence people’s trust, including perception benefits, feelings, knowledge This study aims use these AVs using a survey conducted with 1175 participants. For each participant, 23 features were extracted from questions capture his/her knowledge, perception, experience, behavioral assessment, feelings about These then used as input train an eXtreme Gradient Boosting (XGBoost) With help SHapley Additive exPlanations (SHAP), we able interpret predictions further improve explainability model. Compared traditional regression models black-box machine learning models, our findings show that this approach was powerful providing high level predictability AVs, simultaneously.