A Theory-Based Methodology for Analyzing Domain Suitability for Expert Systems Technology Applications
Abstract
This thesis chronicles the development of a theory-based methodology for analyzing candidate application domains for expert systems technology solution. There is clearly a need for a methodology of selecting expert systems application domains that has firm theoretical underpinnings. To support that goal, four theories of human cognition are evaluated. Gestalt theory, stimulus- response theory, information processing theory, and the Structure of Intellect theory (SOI) are assessed against the criteria of construct validity, reliability, and operational utility. The SOI theory is selected as the framework for identifying the kinds of information and mental processes (the information elements) essential to satisfactory performance in a given domain. The SOI theory forms the basis of the Domain Suitability Analysis Tool (DSAT). By using an established theory of human cognition, the methodology enjoys a high level of construct validity for describing the knowledge components of a domain. The DSAT is hypothesized to specify the information requirements of a domain, present the data in an easily interpreted format, and possess both reliability and validity. The DSAT outputs can then be used to assess the degree of supportability for the domain requirements based upon the current level of expert systems technology. This allows the decision-maker to assess the degree of suitability of this domain for an expert systems technology solution.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1989
- Accession Number
- ADA217648
Entities
People
- Jay A. Horn
Organizations
- Air Force Institute of Technology