Prediction and behavioral analysis of travel mode choice: A comparison of machine learning and logit models

•A comprehensive comparison between machine learning and logit models is provided.•Random forest model has much higher predictive accuracy compared to multinomial logit model and mixed logit model.•Machine learning and logit models agree on many aspects of the behavioral outputs.•Applying a standard approach to generate marginal effects and arc elasticities for random forest lead to unreasonable estimates.•Random forest captures nonlinear relationships between the input features and choice outcomes.