Reply-To: Seethu Seetharaman 
Subject: Ad Model Builder

Hi all: This e-mail is to recount my experience with Dave Fournier's Ad
Model software. I contacted him about the viability of estimating a
Multinomial Probit model on panel data. Once I explained the mathematical
structure of the model to him, he rapidly came up with a code and provided
me instructions on how to convert it into a C++ file and compile and execute
it. Allowing for differences in our programming ability (!), I must say that
his code coverged far more rapidly than mine (in Gauss) and coughed out the
estimates in little time (a homogeneous probit took about 30 minutes to
converge with a 3888*15 predictor matrix that I provided). Looks like his
software is great for DLLs. I am now trying to implement a probit with
correlated random effects using Ad Model Builder. This will involve Monte
Carlo integration (unlike the earler case which employs Gaussian quadrature
for numerical integration) and demonstrate the efficacy of his software for
Monte Carlo-based estimation.

The short of it is that I am very impressed with the Ad Model software and
am adding it to my software tool kit.

Cheers,
Seethu

P.B. Seetharaman
Assistant Professor of Marketing
John M. Olin School of Business
Washington University
Campus Box 1133
One Brookings Drive
St. Louis, MO 63130-4899
Ph: (314)935-5027
web: http://www.olin.wustl.edu/faculty/seetharaman/


.