Printer Nightmare Belief Network from D. Barber "Bayesian Reasoning and Machine Learning"
The demo shows how to:
- Learn model parameters from data (issue and component state history)
- Query trained model
The model is trained from 15 visits of a repairmen to CheapCo. The query infers the probability of Fuse Assembly fault given that Burning smell and Paper Jam is observed (no other issues).
Updated to use open source Infer.NET.