GR-2 Hybrid Knowledge-Based System Using General Rules

GR-2 is a hybrid knowledge-based system consisting of a Multilayer Perceptron and a rule based system for hybrid knowledge representations and reasoning. Knowledge embedded in the trained Multilayer Perceptron (MLP) is extracted in the form of general (production) rules-- a natural format of abstract knowledge representation. The rule extraction method integrates Black-box and Open-box techniques on the MLP, obtaining feature salient and statistical properties of the training pattern set. The extracted general rules are quantified and selected in a rule validation process. Multiple inference facilities such as categorical reasoning, probabilistic reasoning and exceptional reasoning are performed in GR-2. Experiments have shown that GR2 is a reliable and general model for Knowledge Engineering.