MetaProlog: Design, implementation, and application to a medical expert system in acid-base and electrolyte disorders
The first part of the dissertation tackles the issues involved in the design and implementation of the MetaProlog system and shows how it was realized as an extension of the Warren Abstract Machine (WAM). In the process, the notions of theories, viewpoints, and proof trees as first-class objects are discussed and shown to add to the expressive power of the language and make it more attractive for a wider range of Artificial Intelligence applications. The second part deals with a medical expert system in the area of Acid-Base and Electrolyte Disorders which exploits the power and flexibility of the MetaProlog system. It uses a new approach to diagnostic reasoning which can be termed as first-principles assisted evidential reasoning, combined with the concept of progressively expanding diagnostic possibilities. Unlike the first-generation medical expert systems which relied mainly on phenomenological associations, this system incorporates some pathophysiological knowledge to enhance the quality of the diagnosis. But, unlike some new systems that take a multi-level approach and try to establish a diagnosis that is coherent at every level, a more realistic approach to that recognizes the limitations associated with medical knowledge is considered.
- Research Organization:
- Syracuse Univ., NY (USA)
- OSTI ID:
- 7026854
- Resource Relation:
- Other Information: Thesis (Ph. D.)
- Country of Publication:
- United States
- Language:
- English
Similar Records
Automated discovery of medical expert system rules from clinical databases based on rought sets
Temporal reasoning in medical expert systems