Imai, Kosuke and David A. van Dyk. (2005). ``MNP: R Package for Fitting the Multinomial Probit Model,'' Journal of Statistical Software, Vol. 14, No. 3 (May), pp. 1-32.

 

  Abstract

MNP is a publicly available R package that fits the Bayesian multinomial probit model via Markov chain Monte Carlo. The multinomial probit model is often used to analyze the discrete choices made by individuals recorded in survey data. Examples where the multinomial probit model may be useful include the analysis of product choice by consumers in market research and the analysis of candidate or party choice by voters in electoral studies. The MNP software can also fit the model with different choice sets for each individual, and complete or partial individual choice orderings of the available alternatives from the choice set. The estimation is based on the efficient marginal data augmentation algorithm that is developed by Imai and van Dyk (2005) ``A Bayesian Analysis of the Multinomial Probit Model Using the Data Augmentation,'' Journal of Econometrics, Vol. 124, No. 2 (February), pp. 311-334.

  Current Version

The current version is 2.3-3 (built on 06/23/05) which runs on R 2.0 and higher (see what's new? for the explanation of the changes). The source code is available for download from here or The Comprehensive R Archive Network (CRAN). While both articles and software have been published, I hope to keep making improvements to the software in the future. Your omments and suggestions are welcome.

  Installation

Regardless of one's eperating system, the software can be installed by typing the following command at R prompt:

> install.packages("MNP")

This will install the latest version from CRAN.

  Papers

In addition to the Journal of Econometrics article, the accompanying paper (published version; updated version) is available, which includes a brief description of the method implemented in this software as well as a few examples. The software also comes with standard R help files.

  What's New?

VersionRelease dateChanges
2.3-306.23.05made Gibbs sampler slightly more efficient
2.3-206.02.05minor changes to NAMESPACE and DESCRIPTION files
2.3-105.27.05added coef.mnp() and cov.mnp() (thanks to Natasha Zharinova)
2.2-305.12.05minor changes to the documentation; version published in Journal of Statistical Software
2.2-205.09.05minor changes to the documentation
2.2-105.01.05stable release for R 2.1.0; The observations with missing values in X will be deleted in mnp() and predict() (thanks to Natasha Zharinova)
2.1-203.22.05added an option, newdraw, for predict() method
2.1-102.25.05improved predict() method; documentation enhanced and edited
2.0-102.12.05added predict() method (thanks to Xavier Gerard and Saleem Shaik)
1.4-112.16.04improved error handling (thanks to Kjetil Halvorsen)
1.3-211.17.04stable release for R 2.0.1; minor updates of the documentation
1.3-110.09.04stable release for R 2.0.0; updating vector.c
1.2-109.26.04optionally stores the latent variable (thanks to Colin McCulloch)
1.1-209.14.04minor fix in mnp() (thanks to Ken Shultz)
1.1-108.28.04major and minor changes: namespace implemented
1.0-407.14.04users can now interrupt the C process within R (thanks to Kevin Quinn)
1.0-306.30.04bug fix in xmatrix.mnp() (thanks to Andrew Martin)
1.0-206.29.04removed p.alpha0 parameter
1.0-106.23.04official release
0.9-1305.28.04bug fix in ymatrix.mnp()
0.9-1205.23.04updating the documentation and help files
0.9-1105.08.04bug fix in cXnames (thanks to Liming Wang)
0.9-1005.03.04first stable version; bug fix in labeling
0.9-905.02.04bug fix in sampling of W, added summary.mnp() and print.summary.mnp()
0.9-804.29.04improving sampling of W, replace printf() with Rprintf()
0.9-704.27.04missing data allowed for all models, varying choice sets allowed.
0.9-604.26.04missing data allowed in the response variable for standard MNP; a major bug fixed for MoP (thanks to Shigeo Hirano)
0.9-504.25.04improper prior handled by algorithm 1
0.9-404.21.04bug fix in MoP, R-1.9.0 compatible, changes in mprobit.R
0.9-304.10.04rWish() modified with an improved algorithm
0.9-203.22.04first public beta version
0.9-103.20.04first beta version

© Kosuke Imai
Last modified: Wed Aug 3 23:31:03 EDT 2005