Data-Mining Approach to Work Trip Mode Choice Analysis in Chicago, Illinois, Area

Abstract

Discrete choice methods have focused attention on the study of travel mode choice behavior in both theoretical and practical areas of transportation planning and modeling. Unlike discrete choice models that impose predefined probability distributions on choice probabilities, data-mining approaches view the travel mode choice as a pattern recognition problem whereby travel choices can be identified by a combination of explanatory variables. As such, data-mining techniques have enjoyed increasing application in agent-based modeling. The capability of a promising machine learning algorithm, class association rules (CARs), for the estimation of work trip modal choice is examined in the case of the Chicago (Illinois) Area Transportation Studies 1990 Household Travel Survey. The purpose of the study is to investigate the advantages, disadvantages, and applicability of CARs for the study of mode choice behavior. Study results reveal that CARs are useful for building a powerful mode choice model, with the overall accuracy reaching up to 93% for the data set used in this study. Unlike some other mining methods, the rules extracted by CARs are easy to interpret and provide insights into travel behavior from a perspective that is different from that of statistical models. The example presented also illustrates one of the key advantages of CARs over discrete choice models, which is the flexibility of the model specification.