The Asymmetric Triangular Distribution for Continuous Mixed Multinomial Logit Models

A novel method is proposed to estimate a mixed multinomial logit model using the asymmetric triangular distribution as the underlying mixing density. The asymmetric triangular mixing density has the potential to overcome behavioural limitations associated with the most popular mixing densities currently used like the normal, log-normal and Johnson-SB distribution. With only three parameters it is parsimonious, but most important it has a bounded support which can easily be brought in line with behavioural intuitions, i.e. non-positive cost coefficients. Moreover, it is not associated with incredibly large upper (or lower) bounds approaching infinity and it can accommodate varying degrees of skewness in unobserved preference heterogeneity over the population of interest. The proposed method is based on the principle of mixture densities and thereby prevents issues that arise when applying the inverse cdf-method. The asymmetric triangular distribution is relatively flexible in shape and can be applied in a broad range of discrete choice specifications.

• Supplemental Notes:
  • This paper was sponsored by TRB committee ADB40 Transportation Demand Forecasting.
• Corporate Authors:

  Transportation Research Board

  500 Fifth Street, NW
  Washington, DC  United States  20001
• Authors:
  • Dekker, Thijs
• Conference:
  • Transportation Research Board 93rd Annual Meeting
  • Location: Washington DC
  • Date: 2014-1-12 to 2014-1-16
• Date: 2014

Language

• English

Media Info

• Media Type: Digital/other
• Features: References; Tables;
• Pagination: 19p
• Monograph Title: TRB 93rd Annual Meeting Compendium of Papers

Subject/Index Terms

• TRT Terms: Choice models; Maximum likelihood method; Multinomial logits; Probability density functions
• Subject Areas: Planning and Forecasting; Transportation (General); I72: Traffic and Transport Planning;

Filing Info

• Accession Number: 01516616
• Record Type: Publication
• Report/Paper Numbers: 14-3745
• Files: TRIS, TRB, ATRI
• Created Date: Feb 28 2014 1:32PM