The Contribution of Confusion Matrix to the Analysis of Mode Choice in Montreal

Textbooks on travel demand modeling emphasize multinomial logit (MNL) as a pragmatic disaggregate model of mode split. Most often, the key issues are model estimation and model specification including the choice of the explanatory variables, the form of the variables in the utility function, and the definition of the choice set. This paper estimates a simple logit model and then applies a confusion matrix (where the results are partitioned between correct and incorrect predictions) to analyze the errors. The experiment is based upon the Montreal case, where a large regional origin-destination survey was conducted in 1998. The examined dataset consists of 12,383 recorded trips into the Montreal CBD, the center of a region of 3.5 million people. The results indicate that a detailed examination of the residual elements in a logit model can reveal important details about the underlying behavioral phenomena.

• Corporate Authors:

  Transportation Research Board

  500 Fifth Street, NW
  Washington, DC  United States  20001
• Authors:
  • Spurr, Timothy
  • Chapleau, Robert
• Conference:
  • Transportation Research Board 86th Annual Meeting
  • Location: Washington DC, United States
  • Date: 2007-1-21 to 2007-1-25
• Date: 2007

Language

• English

Media Info

• Media Type: CD-ROM
• Features: Figures; Maps; References; Tables;
• Pagination: 15p
• Monograph Title: TRB 86th Annual Meeting Compendium of Papers CD-ROM

Subject/Index Terms

• TRT Terms: Case studies; Central business districts; City planning; Logits; Mode choice; Origin and destination; Travel demand; Travel surveys; Urban areas; Urban development; Urban transportation
• Uncontrolled Terms: Confusion
• Geographic Terms: Montreal (Canada)
• Subject Areas: Planning and Forecasting; Public Transportation; Society;

Filing Info

• Accession Number: 01047512
• Record Type: Publication
• Report/Paper Numbers: 07-2925
• Files: TRIS, TRB
• Created Date: May 2 2007 1:01PM