LRMC5128@BCIT.BITNET (04/28/88)
To: AILIST00--$STRIPE. ailist/ From: Craig Larman Subject: Causal Models and Diagnosis This is my first note - hope it gets through OK. The IntelliCorp (KEE) folks talk enthusiastically about "model based reasoning". They describe this as building a "causal model" of a system (e.g. a gas engine, a liver) and then extending it for various needs (simulation, diagnosis, training). As a programmer, I am interested in programming/technical answers or programming references to these questions: 1. I think of a software "causal model" as defining relationships between components and their dependant inputs/outputs. Is this it? What programming hints or references can folks suggest that will help me learn to build such models? SMALLTALK, LISP, or PROLOG are the preferred languages. 2. Assume I've built a causal model, HOW to build a DIAGNOSIS system from it? I can see simulation follows naturally, but how diagnosis? e-mail address: LRMC5128@BCIT on the NetNorth network (the Canadian arm of BitNet). Virtually, Craig Larman, 432-8629 MIS/Development Centre