Loading...

Private Car Ownership Modeling

Ghodratabadi, Masoud | 2012

731 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 44168 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Kermanshah, Mohammad
  7. Abstract:
  8. Although car ownership models are not one of the four steps of the classical travel demand modeling, an external car ownership model or an internal car ownership sub-model is used in many transport model systems, as an input to mode choice and sometimes also to trip generation and distribution. Utilizing data of Tehran's transportation master plan conducted in 2004, this study develops models for estimating private car ownership with aggregate data (at zonal level) and disaggregate data (at household level) and inspects factors influencing car ownership at both levels. In aggregate modeling, methods of linear regression, geographically weighted regression (GWR) and Quarmby and Bates are used while in disaggregate modeling methods of multinomial Logit, nested Logit and Poisson regression are evaluated. Results of aggregate modeling clarify that, due to geographical distribution and non-linear nature of car ownership modeling, methods such as GWR and Quarmby and Bates surpass the conventional method of linear regression. In disaggregate modeling, for strong theoretical background, Logit models and specifically nested Logit, develop more desirable outcomes than Poisson regression. Results of final models demonstrate that zonal private car ownership boosts with increase in amount of variables such as land value, proportion of people with car driving license, proportion of highly educated people and proportion of people aged fewer than 18 and with decrease in amount of population density. At household level, increase of educational level and proportion of members with car driving license and decrease of household size and proportion of members with motor cycle driving license beside younger structure of a household, lead to growth in probability of owning a private car in a household
  9. Keywords:
  10. Car Ownership ; Linear Regression ; Geographically Wieghted Regression ; Multinomial Logit Model ; Nested Logit Modeling ; Poisson Count Regression

 Digital Object List

 Bookmark

No TOC