THE ESTIMATION OF MULTINOMIAL PROBIT MODELS: A NEW CALIBRATION ALGORITHM

In this paper a Probit estimation procedure is proposed, which uses a numerical approximation to the cumulative multivariate normal developed by Mendell and Elston for the analysis of multifactorial biological traits. It is shown that this procedure provides a more accurate approximation of the Probit multinomial probabilities than the procedure currently used in some Multinomial Probit estimation algorithms, leading to more accurate estimates of the model parameters. This procedure is presented after a discussion of other approaches. The choice probabilities resulting from this approximation are then compared with the ones obtained from other methods. Finally, a Probit estimation algorithm, using the proposed approximation, is tested.

• Availability:
  • Find a library where document is available. Order URL: https://www.library.northwestern.edu/find-borrow-request/requests-interlibrary-loan/lending-institutions.html
• Authors:
  • KAMAKURA, W A
• Publication Date: 1989-11

Media Info

• Features: Figures; References; Tables;
• Pagination: p. 253-265
• Serial:
  • Transportation Science
  • Issue Number: 4
  • Publisher: Institute for Operations Research and the Management Sciences (INFORMS)
  • ISSN: 0041-1655
  • Serial URL: http://transci.journal.informs.org/

Subject/Index Terms

• TRT Terms: Algorithms; Calibration; Choice models; Forecasting; Multinomial probits; Multinomials; Probits
• Old TRIS Terms: Human choice behavior
• Subject Areas: Highways; Planning and Forecasting; I72: Traffic and Transport Planning;

Filing Info

• Accession Number: 00490618
• Record Type: Publication
• Files: TRIS, ATRI
• Created Date: Jan 31 1990 12:00AM