In this paper the applicability of an artificial neural network for modelling modal choice is examined. Characteristics of transportation modes and of individuals are used as input variables, and the mode selected is the output variable. We applied the network to the Indian city of Guwahati by using various numbers of modes. Although the trained network reproduced the observations in the training data accurately, predictability was low when the untrained data set was used. However, predictabilty was better when a smaller number of modes were modelled.
Get full access to this article
View all access options for this article.
References
1.
FaghriAHuaJ, 1992, “Evaluation of artificial neural network applications in transportation engineering”Transport Research Record135871–80
2.
KungS Y, 1993Digital Neural Networks (Prentice-Hall, Englewood Cliffs, NJ)
3.
LippmanP R, 1987, “An introduction to computing with neural nets”IEEE ASSP Magazine44–22
4.
ShimazakiT, 1992, “Model choice model by neural network theory”; copy available from the author
5.
TaylorJ G, 1993The Promise of Neural Networks (Springer, Berlin)
6.
WassermanP D, 1993Advanced Methods in Neural Computing (Van Nostrand Reinhold, New York)
