Abstrakt: |
Ontological patterns are a tool for reducing the labor intensity of developing intelligent software systems. The paper presents a pattern for a class of diagnostic tasks. It contains a complex of domain-independent semantic structures for the formation of knowledge, input documents, diagnostic results, and its explanation. The complex of structures is supplemented by means of their adaptation to specialized domain terms. The universal diagnostic solver works with knowledge and other information generated in accordance with these semantic structures. The result of its reasoning is a set of unrejected versions of the diagnosis (malfunction, anomalous process, deviation). Thanks to the new adaptation method, knowledge and data are formed and viewed in the terminology (and markup) of the domain. At the same time, problem-oriented (diagnostic) logical inference does not depend on these particular terms. Therefore, diagnostic software systems for various practical fields of activity are effectively produced on the basis of a single pattern and solver. [ABSTRACT FROM AUTHOR] |