Providing a Revenue-forecasting Scheme to Relocate Groups of Ride-Sourcing Drivers

Abstract

The selection of transportation modes has been a popular topic in transportation planning, and it has been studied for several decades using random utility optimization. However, recent developments in machine learning techniques have opened up new opportunities for accurate prediction. In this study, we used data from the Chicago including activity and travel records for 12,000 households in a 24-hour period. Our objective was to develop travel mode choice models for commute trips through estimating a multinomial logit model to identify the factors that influence people's choices of commute travel mode. Notably, our study is the first in Chicago to consider seven travel modes, including walking, biking, walking to transit, driving to transit, auto driver, auto passenger, and TNC. Additionally, we employed a machine learning classifier to model the mode choice problem and compared its performance with the econometric model.