Modeling mode choice preference in a Mexican university with discrete choice models.

The study of the mode choice for urban regions has increased, with a growing set of recent works using joint methodologies of data gathering and modeling with machine learning models. This work details the design and application of a mobility survey to a private urban university in the north of Mexico. Decision tree based, machine learning models for multiclass classification, are shown to be effective with datasets in which categorical data predominates, having a better performance than the widely applied econometric models covered in literature. The interpretability of decision trees helps to identify relevant variables that influence modal choice. It can be concluded that, for the studied sample, people's awareness of their access to collective modes is the most decisive factor, and thus the efforts of institutions to promote investments and availability of better modes will determine the mode's adoption rate. [ABSTRACT FROM AUTHOR]