ABSTRACT
As a social science tool, neural network analysis is still in the initial stage. This chapter demonstrates the benefits and limitations of applying the technique to a social science problem in this case the prediction of travel behaviour as a necessary preliminary to transportation planning. In order to test the efficacy of the neural network approach, it was necessary to identify a body of data that provided complex information with numerous and specific factors. An appropriate combination was selected, consisting of the Israeli Travelling Habits Survey for 1984 and data from the 1983 Israeli Population and Housing Census. The neural network model structure developed for this experiment is outlined. Hypotheses and results of selected neural network experiments are presented, along with a discussion of the benefits and limitations, as well as recommendations for future research in the use of neural network analysis in predicting a behavioural outcome.
