Open Collections

Comparison of neural classifiers and conventional approaches to mode choice analysis

University of British Columbia

This thesis provides a comparison of three modeling techniques which can be used for mode choice analysis. The techniques include the conventional logit, artificial neural networks (ANNs), and neurofuzzy models. The three modeling techniques were applied to mode choice data extracted from the 1999 24-hour trip diary survey of the Greater Vancouver Regional District. The travel mode of each individual was explained using explanatory variables acquired from three categories of the database: household database, personal database, and trip database. The results showed that, as modeling techniques, both ANNs and neurofuzzy models are highly adaptive and very efficient in dealing with problems involving complex interrelationships among many variables. The neurofuzzy technique combines the learning ability of artificial neural networks and the transparent nature of fuzzy logic. In addition; the neurofuzzy technique only selects the variables that significantly influence mode choice and display the stored knowledge in terms of fuzzy linguistic rules. This allows the modal decision making process to be examined and understood in great detail. The results of the comparison also indicated that neurofuzzy models produced the best results in terms of model accuracy. As well, it selected the least number of variables to achieve these results.

Thesis/Dissertation

application/pdf

2009-08-12

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.

http://hdl.handle.net/2429/12050

Civil Engineering

University of British Columbia

UBCV

DSpace