A comparison of the predictive potential of artificial neural networks and nested logit models for commuter mode choice

Understanding and predicting traveller behaviour remains a complex activity. The set of tools in common use by practitioners and many of the tools used by researchers appear in many ways to exhibit complexity; yet often this richness of detail is in methods of estimation rather than in representation of how individuals actually evaluate alternatives and make decisions on a set of interrelated travel choices. Discrete choice methods championed by the multinomial logit model and its variants such as nested logic, heteroskedastic extreme value, and multinomial probit have added substantial behavioural richness into statistical specification and estimation (Hensher et al forthcoming), seeking to accommodate the role of both observed and unobserved influences on travel choices. The search for behavioural and analytical enhancement continues. Research in the field of artificial intelligence systems has been exploring the use of neural networks (eg Faghri and Hua 1991, Yang et al 1993) as a framework within which many traffic and transport problems can be studied. The main motivation for using neural networks could be due to some fascinating properties that neural networks possess. They are parallelism, the capacity to learn, allowing for the use of distributed memory and capacity for generalisation. Following these characteristics, one of the promises from neural networks is that they can tackle the problem of forecasting and modelling which is very common in travel demand modelling. The use of such tools in studying individual traveller behaviour opens up an opportunity to consider the extent to which there are representation frameworks which complement and/or replace existing analytical approaches. This paper explores the merits of neural networks as part of a revised framework within which to explore the processes of traveller decision making, and how discrete choice methods might be integrated within such a framework to acknowledge the important role that the latter tools have played in the last 25 years in the development of better practice in travel demand modelling.