• PAPER:  "The Maximum Approximate Composite Marginal Likelihood (MACML) Estimation of Multinomial Probit-Based Unordered Response Choice Models" by C.R. Bhat  PDF version, MS Word version

If you use any of the GAUSS or R codes (in part or in the whole/ rewrite one or more codes in part or in the whole to some other language), please acknowledge so in your work and cite the paper listed above as well as the following paper:  

Bhat, C.R., and R. Sidharthan (2011b), "A Simulation Evaluation of the Maximum Approximate Composite Marginal Likelihood (MACML) Estimator for Mixed Multinomial Probit Models", Transportation Research Part B, 45 (7), 940-953.

• The Maximum Approximate Composite Marginal Likelihood (MACML) - GAUSS Codes:    GAUSS documentation, codes, test data sets, and estimation results
  
  
  • Download Here (zip file)
    This zip file contains the following four items:
    1. A file documenting the MVNCD and MACML codes
    2. A folder labeled "codes" with the actual Gauss codes
    3. A folder labeled "data" with the sample data sets used by the codes
    4. A file with the estimation results for the sample data sets

     

• Second Order Approximation of MVNCD function - GAUSS Codes:    GAUSS code, documentation, and test data sets
  
  
  • Download Here (zip file)
    This zip file contains the following items:
    1. A folder labeled codes containing GAUSS implementation of the second order(SO) approximation of MVNCD function
    2. Documentation for second order MVNCD approximation.docx
    3. Calling SO.gau - A sample GAUSS code demonstrating how the SO functions are to be invoked

     

• The Maximum Approximate Composite Marginal Likelihood (MACML) - R Codes:    R code, documentation, and test data sets
  
  
  • Download Here (zip file)
    This zip file contains the following four items:
    1. A file documenting the MVNCD and MACML codes
    2. A folder labeled "codes" with R codes
    3. A folder labeled "data" with sample data sets used by the codes
    4. A file - "results.txt" - with the estimation results for the sample data sets

     

• Return to list of codes