Robert McCulloch

Professor of Econometrics and Statistics
University of Chicago
Graduate School of Business
5807 South Woodlawn Avenue
Chicago, IL, 60637

email: robert.mcculloch@chicagogsb.edu

phone: (773) 702-7315 fax: (773) 702-0458


photo of Robert McCulloch circa 2000

My research is in Applied Bayesian Statistics.

teaching applets

CV    (some recent papers are here)

Some recent talks:

Fully Nonparametric Bayesian Ensemble Modeling
(JSM, July 2007)

General Equilibrium Options Pricing Under Recursive Preferences
(Iowa time series meeting, September 2007)

Bayesian Ensemble Modeling with Application to Causal Inference
(University of Texas, September 2007)

BayesTree: An R Package for Flexible Bayesian Inference
(New Mexico, October 2007)

Choleski Multivariate Stochastic Volatility
(given by Ruey Tsay at Colorado time series meeting, June 2007)





BayesTree:

Below is the tarball for the latest version of the R-package "BayesTree" (0.3-0).
which implements BART: Bayesian Additive Regression Trees by Chipman, George, and McCulloch
(see CV link above for papers).
The tarball was built using R-2.4.0.

(Version 0.2-0 is on CRAN)

WBayesTree_0.3-0.tar.gz

As usual, to install (under Linux): $ R CMD INSTALL WBayesTree_0.3-0.tar.gz

Here is a R-script which runs some simulated examples with binary y:

test_bin.R

Feedback very welcome of course.

The 0.4-0 version is "very beta" , it has the Dirichelt Process prior on the errors.

WBayesTree_0.4-0.tar.gz





Hill and McCulloch ``Bayesian Nonparametric Modeling for Causal Inference'',
simple R script (simple_do-lalonde.R) and data files (nswre74_treated.txt, psid_controls.txt)
for Lalonde example in paper.



Old CART code and some variable selection code is here.



for dru:

draws using options data:

pdr_md.dat    

columns are 'sigma', 'sigma_k','sigma_t','rho','delta','tbar','gamma','psi','beta','mu'


thetadr_md.dat  

each row is a draw of the state.


rows in pdr_md.dat should be from the same draw as the corresponding row of thetadr_md.dat




first data set (72 time periods):

pdr1.dat   thetadr1.dat

second data set (60 time periods):

prd2.dat   thetadr2.dat

thetadr_cons0.dat    





for_ron.tgz