Paper Title
Understanding Mode Choice Decision Making of Commuters Using Different Modelling Framework

Abstract
This study examines decision making behaviour of 1350 households from low and low-middle income brackets living in Bengaluru, India regarding mode choice classification using ML classifiers like Decision Trees, Random Forest, Extreme Gradient Boosting and Support Vector Machines. Random forest performed the best (0.61 on training data and 0.4 on testing data) among all models. This research has adopted interpretability techniques like feature importance to explain the decision making behaviour using ML models. Key findings indicate that travel cost and travel time contribute highest to the model’s predictive output, across all ML models.