Functions for Hierarchical Bayesian Estimation: A Flexible Approach

Functions for estimating models using a Hierarchical Bayesian (HB) framework. The flexibility comes in allowing the user to specify the likelihood function directly instead of assuming predetermined model structures. Types of models that can be estimated with this code include the family of discrete choice models (Multinomial Logit, Mixed Logit, Nested Logit, Error Components Logit and Latent Class) as well ordered response models like ordered probit and ordered logit. In addition, the package allows for flexibility in specifying parameters as either fixed (non-varying across individuals) or random with continuous distributions. Parameter distributions supported include normal, positive/negative log-normal, positive/negative censored normal, and the Johnson SB distribution. Kenneth Train's Matlab and Gauss code for doing Hierarchical Bayesian estimation has served as the basis for a few of the functions included in this package. These Matlab/Gauss functions have been rewritten to be optimized within R. Considerable code has been added to increase the flexibility and usability of the code base. Train's original Gauss and Matlab code can be found here: < http://elsa.berkeley.edu/Software/abstracts/train1006mxlhb.html> See Train's chapter on HB in Discrete Choice with Simulation here: < http://elsa.berkeley.edu/books/choice2.html>; and his paper on using HB with non-normal distributions here: < http://eml.berkeley.edu//~train/trainsonnier.pdf>. The authors would also like to thank the invaluable contributions of Stephane Hess and the Choice Modelling Centre: < https://cmc.leeds.ac.uk/>.


News

RSGHB 1.2.0

  • Added outputs to allow for better integration with CMC model estimation code
  • Added additional options for sampling from posterior of the covariance matrix - hierarchical Inverted Wishart

RSGHB 1.1.2

  • Fixed a bug in writeModel when only one random parameter is specified

RSGHB 1.1.1

  • Fixed a bug when estimating a model with thinning (gNSKIP > 1)

RSGHB 1.1.0

  • doHB now defaults to normal distribution assumptions, if not specified
  • doHB now defaults to starting values of 0, if not specified
  • RSGHB has been made R-object oriented rather than CSV-file oriented
  • This primarily affects how users interact with model outputs
  • doHB now returns a model object of class 'RSGHB'
  • Save results to disk by setting the 'writeModel' argument in doHB to TRUE, or by using the new 'writeModel' function on an 'RSGHB' object
  • Updated documentation, examples, and vignette
  • General code improvements to increase speed and reduce memory footprint

RSGHB 1.0.2

  • Added ability to fix the mean and variance of random parameters
  • plotC function added to plot the posterior means

RSGHB 1.0.1

  • General updates and bug fixes

RSGHB 1.0.0

  • Functions for estimating Hierarchical Bayesian models

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

install.packages("RSGHB")

1.2.1 by Jeff Dumont, 15 days ago


https://github.com/RSGInc/RSGHB


Browse source code at https://github.com/cran/RSGHB


Authors: Jeff Dumont [aut, cre] , Jeff Keller [aut] , Chase Carpenter [ctb]


Documentation:   PDF Manual  


Task views: Bayesian Inference, Econometrics


GPL-3 license


Imports graphics, grDevices, stats, utils

Depends on methods, MCMCpack


Imported by apollo.


See at CRAN