Public Transport Route Choice Model Based On Feature Importance Derived From Clustering Analysis

Abstract

This paper examines the importance of public transport’s unique characteristics in the estimation of public transport route choice models by applying clustering analysis to identify significant patterns in the sampled data. The proposed methodology employs a data-driven approach for generating the choice set and for characterizing the important explanatory variables in the route choice model. The utility specification of the public transport route choice model is derived from the feature importance obtained in the clustering results. The feature selection based on the clustering yields significant explanatory variables in the public transport route choice model.