Abstract This paper describes the results obtained in applying the learning system ENIGMA to a fault diagnosis problem of electromechanical devices at ENICHEM (Ravenna, Italy). The system ENIGMA is capable of learning structured knowledge from examples and a domain theory, using an integrated inductive/deductive paradigm The results are compared with the ones obtained by an expert system, designed for the same task, in which the knowledge base was acquired using the traditional method of expert interview. The comparison indicates that performances obtained by the learning system are systematically better than the ones obtained by the manually developed expert system. The conclusion is that, even if still leaving room for improvements, automated learning is a viable approach to the construction of expert systems, from the point of view of both obtainable performance and of limiting the development time and cost.
[1]
Stephen Muggleton,et al.
Machine Invention of First Order Predicates by Inverting Resolution
,
1988,
ML.
[2]
Lorenza Saitta,et al.
Automated Concept Acquisition in Noisy Environments
,
1988,
IEEE Trans. Pattern Anal. Mach. Intell..
[3]
Ivan Bratko,et al.
ASSISTANT 86: A Knowledge-Elicitation Tool for Sophisticated Users
,
1987,
EWSL.
[4]
Ryszard S. Michalski,et al.
A Theory and Methodology of Inductive Learning
,
1983,
Artificial Intelligence.
[5]
Ryszard S. Michalski,et al.
The AQ15 Inductive Learning System: An Overview and Experiments
,
1986
.
[6]
Francesco Bergadano,et al.
A Knowledge Intensive Approach to Concept Induction
,
1988,
ML Workshop.