Bayesian Approaches to Learning from Data how to Untangle the Travel Behavior and Land Use Relationships
dc.contributor.advisor | Clifton, Kelly J | en_US |
dc.contributor.author | Scuderi, Marco Giovanni | en_US |
dc.contributor.department | Urban and Regional Planning and Design | en_US |
dc.contributor.publisher | Digital Repository at the University of Maryland | en_US |
dc.contributor.publisher | University of Maryland (College Park, Md.) | en_US |
dc.date.accessioned | 2006-02-04T07:34:24Z | |
dc.date.available | 2006-02-04T07:34:24Z | |
dc.date.issued | 2005-12-05 | en_US |
dc.description.abstract | The body of research on land use and travel behavior relationships reaches widely different conclusions with results varying even when considering a single author. The hypothesis of this research is that these differences arise, in part, from the fact that the vast majority of these studies do not address all the theoretical travel behavior tenets and are therefore ad-hoc in nature. An inductive approach to the study of the relationships between land use and travel behavior, prior to carrying out traditional deductive studies, can help improve the outcomes by providing an opportunity to identify and test such relationships. With data sourced from the 2001 National Household Travel Survey Add-On, supplemented with local land use data, this study uses heuristic search algorithms to evaluate relationships hidden in the data without these being framed, a priori, by specific statistical constructs. Bayesian scoring is used to evaluate and compare the results from actual data collected for the Baltimore Metropolitan Area with the set of predominant conceptual frameworks linking travel behavior and land use obtained from the literature. Results show that socioeconomic factors and land use characteristics act in a nested fashion, one in which socioeconomic factors do not influence travel behavior independently of land use characteristics. The land use travel behavior connection is specifically strong only for particular combinations of socioeconomic characteristics and a land use mix which includes both moderate residential densities and a significant amount of commercial opportunities. The study also finds that the heuristic search approach to derive relationships between land use and travel behavior does work, that this technique needs to be fine tuned for the proper use of spatially explicit data, and that although the research outputs are an unbiased representation of the land use travel behavior relationships, they need proper interpretation, especially in light of persisting theoretical questions still driving this research field. The study concludes that an inductive approach to the analysis of the relationships between land use and travel behavior provides valuable knowledge of the data that can be used to better formulate deductive studies, so that the two methodologies are complementary to each other. | en_US |
dc.format.extent | 2614291 bytes | |
dc.format.mimetype | application/pdf | |
dc.identifier.uri | http://hdl.handle.net/1903/3201 | |
dc.language.iso | en_US | |
dc.subject.pqcontrolled | Urban and Regional Planning | en_US |
dc.subject.pqcontrolled | Transportation | en_US |
dc.title | Bayesian Approaches to Learning from Data how to Untangle the Travel Behavior and Land Use Relationships | en_US |
dc.type | Dissertation | en_US |
Files
Original bundle
1 - 1 of 1