Methods of Eliciting Information from Experts
Abstract
The biggest bottleneck in the development of expert systems is the problem of eliciting from experts the mechanisms responsible for their expertise. This report examines what is known about experts and suggests a number of ways of eliciting information from them. The literature suggests that the mechanisms of expertise represent deep-seated ways of conceptualizing and perceiving stimuli, and that these mechanisms must be differentiated from relatively superficial proceduralized rules, which make up most of what expert information consists of today. The goal of reproducing the expert's mental processes in a computer system appears unrealistic at present. The only way of determining that one has, in fact, tapped expertise is to build the expert system and evaluate its effectiveness. If it matches or surpasses human proficiency, one has incorporated human expertise into the system. Keywords: Human expertise; Knowledge acquisition; Mental processes; System design; Data acquisition.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1987
- Accession Number
- ADA187468
Entities
People
- David Meister