Evaluating expert systems by formal metrics
Zhisong Chen, Ching Y. Suen · 2002
Expert systems (ES) have been used as sophisticated tools to solve complex problems in many areas, such as fault diagnosis in telecommunication and reactive power and voltage control in power electronics. But, few adequate quantitative measures of ES exist, this leads to risks in applying ES. In every discipline, measures are essential. ES, too, needs such measures for assessment. We investigate some characteristics of ES, and present and describe several expert system metrics, including our proposed complexity measure (the RC metric), to formally measure these characteristics.>