Construction of diagnosis rules from diagnosis table based on error risk criterion

A diagnosis table, which shows the relation between symptoms and abnormal events to be diagnosed, is a compact expression of diagnosis knowledge. Since derivation of diagnosis rules from a diagnosis table is conventionally based on only its logical structure, some parts which have identical symptoms with different abnormal events are omitted in their construction. Some cases corresponding to these cases cannot be dealt with by the established rules. To solve this problem, this paper proposes a construction method of diagnosis rules based on an error risk criterion. This criterion can add the system analyst´s evaluation of the results as well as the prior information on abnormal events to the established rules. An illustrative example shows the details and merits of the proposed method