Knowledge Extraction Through Learning from Examples
We present a method for a knowledge-base compression, restricted to rules of specific form, which uses a learning from examples facility. A large medical knowledge-base, which was a source of more than 5000 training examples, was compressed to 3% of its original size. Most of the extracted rules turned out to be quite meaningful from the medical point of view.
KeywordsDecision Variable Disjunctive Normal Form Decision Class Inductive Learning Heart Disorder
Unable to display preview. Download preview PDF.