Note that there are at least three versions of the ALARM network floating around. They have the same structure, but different probabilities: 1. Used in Cooper and Herskovits, Machine Learning, 1995. 2. A later version used in Heckerman, Geiger, Chickering, Machine Learning, 1997. This net is available at ftp.research.microsoft.com/pub/dtg/alarm/alarm.dsc. 3. An even later version that is included with Herskovits' Bayes-net tool called Ergo. I'm not sure which one of these (if any) is on the repository. David -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Friday, March 10, 2000 1:55 PM Subject: Hi - The ALARM network was originally developed by Beinlich et al. as an experimental prototype that helps interpret monitoring data to alert anesthesiologist to various situations in the operating room. It has about 37 nodes, some binary, some non-binary. The original paper about ALARM is the following: http://smi-web.stanford.edu/pubs/SMI_Abstracts/SMI-88-0239.html I. A. Beinlich, H. J. Suermondt, R. M. Chavez, & G. F. Cooper. The ALARM Monitoring System: A Case Study with Two Probablistic Inference Techniques for Belief Networks. Second European Conference on Artificial Intelligence in Medicine, London, 38:247-256. 1989. ALARM now ships as an example network with most common inference engines. You can download it from the Bayesian network repository at Berkeley - http://www-nt.cs.berkeley.edu/home/nir/public_html/Repository/ ALARM was first used in belief-network learning by Herskovits and Cooper as a test of their learning methods, Kutato and K2. An early paper on that is: http://smi-web.stanford.edu/pubs/SMI_Abstracts/SMI-90-0305.html E. Herskovits & G. Cooper. Kutato: An Entropy-Driven System for Construction of Probabilistic Expert Systems from Databases. Sixth Conference on Uncertainty in Artificial Intelligence, Cambridge, MA, 54-62. 1990. Hope this helps Jaap Suermondt Hewlett-Packard Laboratories, Palo Alto, CA YueBo wrote: > > Hello all, > As an example, many people who devote learning > > Bayesian networks use the "Alarm network" to > illustrate their algorithms. Now I am working at > learning Bayesian networks and could anyone tell > me where I can get the Alarm network? > Thanks a lot! > > -- > YUE Bo, Ph.D. > Key Lab for Radar Signal Processing > Xidian University > Xi'an, 710071 > P.R.China > > Email: [EMAIL PROTECTED] ------- End of Forwarded Message
