An automatic segmentation procedure for studying variations in mode choice behaviour

摘要: Urban areas are very complex and heterogeneous in terms of their population composition activity systems. The transport system, modal choices service levels available to the also varies considerably across space time. These similarities differences explanatory variables facing individual tripmakers have be explicitly considered any study behvior. common practice has been include user attributes, addition system characteristics, utility functions help capture choice behavior individuals. However, it could well that mode-choice a segment is fundamentally different from other segments population. In view this, some studies applied segmentation schemes identify subgroups presumably travel responses. Typically, such based on stratification by single variable, chosen either priori notions or one-way cross tabulations. shortcomings. Thus, this paper develops an analytical procedure simultaneously deals with level service, socioeconomic spatial factors determine relative role each plays determining behavior. data Toronto region illustrate its use.