Predicting Travel Mode of Individuals by Machine Learning

Travel mode choice prediction of individuals is important in planning new transportation projects. In this paper, we present four machine learning methods namely artificial neural net-MLP, artificial neural net-RBF, multinomial logistic regression, and support vector machines, for predicting travel mode of individuals in city of Luxembourg. The presented methods use individuals’ characteristics, transport mode specifications and data related to places of work and residence. The dataset analyzed comes from a national survey. It contains information on the daily mobility (e.g., from home to work) of individuals who either live or work in Luxembourg. We extracted individual characteristics to relate daily movements (journeys between home and work, in particular) to the characteristics of working individuals. We used the information about public transportation and some geographical location of the residential and work places. We compare the rates of successful prediction obtained by neural networks and several alternative approaches for predicting the travel mode choice using cross-validation. The results show that the artificial neural networks perform better compared to other alternatives. Our analysis can be used to support management decision-making and build predictions under uncertainty related to changes in people's behavior, economic context or environment and transportation infrastructure.