Predicting Mode Choice through Multivariate Recursive Partitioning

Abstract

Understanding and predicting individual mode choice decisions can help address issues ranging from forecasting demand for new modes of transport to understanding the underlying traveler behavior and characteristics. Early research in mode choice modeling revolved, almost exclusively, around the family of logit models. But a number of researchers have recently argued that these models place restrictions on their parameters that compromise their performance and have thus experimented with a number of newly developed, flexible mathematical techniques. The present paper extends prior research by developing a methodology for predicting individual mode choice based on a nonparametric classification methodology that imposes very few constraining assumptions in yielding mode choice predictions. Preliminary results, using data from three vastly different international settings, are promising, especially when considering that the models are successful while using only a limited number of independent variables to achieve these predictions.