碩士 / 國立成功大學 / 交通管理(科學)學系 / 85 / ABSTRACT This research discussed
the temporal transferability of intercity disaggregate mode
choice models by the updating techniques and model fitness. The
intercity travel data were collected in the Taipei-Tainan
corridor between 1994 to 1997 for the empirical study. Beside
provides comparison of four updating techniques within different
model specification, this research using combined transfer
estimation to compare the model transferability within different
utility models. At the end, the sensitivity analysis is applied
to discuss model transfer performance. The empirical results
showed:1.Model specification affects the perfomance of updating
techniques and updating effect.2.Among the four updatings
techniques, joint context estimation and combined transfer
estimation have a better performance than others. For further
measurement, it indicate that joint context estimation is the
most effective updating method.3.There''s no difference in
transfer effectiveness between earlier and later updating
models.4.The indifference threshold models and the tradtional
multinomial logit models are transferable when using combined
transfer estimation. Every measures indicate the former models
are better than the latter ones.5.The sensitivity analysis
indicate all updating models can correctly reflect the changes
of the market share. By the percentage of aggregate prediction
error measure, the combinded transfer estimation and joint
context estimation are superior to others. Some models such as
Bayesian updating model which have bad performance in
statistics may become better in aggregate forecasting.
| Identifer | oai:union.ndltd.org:TW/085NCKU0119021 |
| Date | January 1997 |
| Creators | Hsieh, Iuan Jiun, 謝淵鈞 |
| Contributors | Cheng-Chih Chang, 張淳智 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 79 |
Page generated in 0.0014 seconds