Expert system

An expert system is a program that uses stored domain knowledge and explicit rules to solve problems or give recommendations in a specialized field.

Knowledge base: A structured store of facts, rules, or other information used by a computer system. It holds the domain facts and rules an expert system applies.
Production system: A computational model built from condition-action rules, working memory, and a control strategy. It is the general architecture underlying many rule-based expert systems.
DENDRAL: An early expert system that inferred molecular structures from mass spectrometry data. It demonstrated how specialized chemical knowledge could guide rule-based inference.
XCON: A rule-based system used by Digital Equipment Corporation to configure VAX computer orders. It showed that an expert system could handle a costly, large-scale industrial task.
Machine learning: A field of methods that enable computers to improve performance by learning patterns from data. Unlike hand-coded expert-system rules, many machine-learning models derive behavior from examples.
Inference engine: A software component that derives conclusions from facts and rules. It applies the system’s rules to known facts to reach conclusions.
Rule-based system: A computer system that represents knowledge as rules and uses them to derive actions or conclusions. Expert systems are a specialized form of rule-based system.
MYCIN: A 1970s expert system that recommended treatments for blood infections using rules and certainty factors. It became a defining example of medical diagnosis and explanation by rules.
Medical diagnosis: The process of identifying a disease or condition from signs, symptoms, and other evidence. Diagnosis was a prominent target for systems such as MYCIN.
Artificial neural network: A computing model of interconnected units that transforms inputs through weighted connections. Its learned distributed representations differ from an expert system’s explicit rules.

Linked from 11 pages