--On Thursday, October 14, 1999, 5:02 PM -0700 Chien-wen Shen
<[EMAIL PROTECTED]> wrote:
> I'm a graduate student looking for the QMR or QMR-DT database for my
> research. Can you help me to find out where and how to get the QMR
> knowledge base or related data? I've tried to contact the University of
> Pittsburgh, Camdat Corporation/First Databank (QMR is licensed to this
> company), and Professor Randy Miller (one of the QMR developers) for
> further information, but I didn't get positive responses from them. We
> might not need to have the whole 600 diseases and 4000 findings knowledge
> base. Only a portion of it (like 60 diseases and 400 findings) may be
> enough for my research. It might even be enough to just know the
> *structure* of the underlying Bayes net (where the arcs are and how many).
> I really appreciate if any one can provide more information about getting
> a copy or partial copy of QMR (QMR-DT) for academic research.
Chien-wen,
The official way is to ask the lawyers at the University of Pittsburgh
Office of Technology Transfer and Intellectual Property. The address is:
University of Pittsburgh
Office of Technology Transfer and Intellectual Property
911 William Pitt Union
Pittsburgh, PA 15260
Telephone: 412-648-2206
FAX: 412-648-8525
I dealt with Frances J. Connell. I would like to warn you that it took me
1.5 years to get permission from them and I work at the University of
Pittsburgh! I did not press really hard, though, and I may have also
created a precedent. Also, they will require you to clear every
publication based on this model through their office before submitting. I
have promptly always disregarded this requirement in my work, as I find it
completely ridiculous. On the other hand, I have always done very
"innocent" things with the model -- I have used it for testing Bayesian
network algorithms.
Good luck,
Marek
--------------------------------------------------------------------------
Marek J. Druzdzel http://www.pitt.edu/~druzdzel