The transferability of travel demand models : an analysis of transfer methods, data quality and model estimation
No Thumbnail Available
URL
Journal Title
Journal ISSN
Volume Title
Doctoral thesis (monograph)
Checking the digitized thesis and permission for publishing
Instructions for the author
Instructions for the author
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Authors
Date
2003-12-17
Major/Subject
Mcode
Degree programme
Language
en
Pages
214
Series
Julkaisu / Teknillinen korkeakoulu, liikennetekniikka, 106
Abstract
The main goal of this study was to compare alternative methods of spatial transfer as a function of sample size, and identify the factors affecting the models quality and the impreciseness of the model parameters. In addition, different test measures for studying model transferability were compared and the applicability of the traditional statistical tests, with respect to those based on the prediction accuracy of sample enumeration tests and forecasts, were assessed. The research primarily concerned the transferability of mode and destination models; however, the preciseness of the trip generation level was considered as well. The study was mainly based on the mobility surveys conducted in the Helsinki Metropolitan Area (HMA) in 1995 and in the Turku region in 1997. The transferring procedures examined were Bayesian updating, combined transfer estimation, transfer scaling, and joint context estimation procedures. The trip groups studied were home-based work trips and other home-based trips. The studied modes were walk and bicycle, car and public transport. To explore the impact of sample size on transferring performance, model transferability was tested using three to four different sample sizes. Thus, all the transferability tests were made by using 100 bootstrap samples (resampled from the Turku 1997 dataset) for each trip group, transfer method and sample size category. The results indicated that joint context estimation gives the best prediction performance in almost all cases. In particular, the method is useful if the transfer bias is large or only some of the coefficients are precise. The applicability of joint context estimation can be improved by viewing the coefficients as variable-oriented and emphasizing precise and imprecise coefficients differently. The models transferred by using combined transfer estimation or transfer scaling were most sensitive to the sample size and their use, therefore, requires much larger samples than the Bayesian approach or joint context estimation. In addition, note that due to repeated measurements the results based on the Bayesian method and combined transfer estimation may be strongly biased. When defining the sample size required the fact that defining mode shares precisely may require more observations than the transferring mode and the destination choice models must be taken into account. The results also showed that statistical tests are not able to evaluate the goodness of transferred models with a high enough degree of versatility. For example two models that have totally different values for coefficients may have the same TTS. As a result, their ability to predict the effect of changes in a transportation system may differ greatly. On the whole, the differences between the best transfer methods are, in some cases, rather small, and the errors caused by the factors connected to the modelling and sample size seem to be larger than the errors caused by the model transfer itself.Description
Keywords
transferability, travel demand model, mode choice, logit model
Other note
Parts
- Additional errata file available.