Методика разработки экспертных систем с использованием графовой СУБД и онтологического подхода.

Abstract

Expert systems are a complex software system that accumulates the knowledge of specialists in specific subject areas and disseminates this empirical experience for consultations with less qualified users, which reduces the number of medical errors caused, in particular, by a lack of specialist doctors in local areas and the time required for a single consultation. P u r p o s e of the study is to present a modern methodology for developing expert systems based on knowledge engineering using an ontological approach and graph databases. M a t e r i a l s a n d m e t h o d s . System analysis methods were applied during the creation of the methodology for developing expert systems, including analytical and descriptive methods, knowledge engineering methods such as structuring methods, knowledge representation, and fuzzy logic. The methodology assumes the use of a graph database as a storage for the knowledge base. R e s u l t s . The methodology describes the process of developing expert systems based on a graph database using an ontological approach. It includes the following components: a knowledge base, an intelligent editor of the knowledge base, a problem solver, and explanation subsystem. The methodology assumes the participation of a group of experts in the development of the expert system, with whom the development of two artifacts is carried out: a terminological database of the subject area and intermediate information objects describing the logic of solving the problem. Both artifacts should be formed automatically using the intelligent solver of the knowledge base, as well as filling the knowledge base, which will facilitate the subsequent maintenance of the knowledge base and keep it up-todate. According to the methodology, the solver and explanation subsystem work with a graph database, which is a means of storing the knowledge base, in order to find the optimal path in which nodes are used to solve the problem of the expert system, for example, supporting a specialist doctor in establishing a preliminary/final diagnosis, and the path is used by the explanation subsystem to argue the proposed solution of the expert system. F i n d i n g s . A methodology for creating expert systems using a graph database and an ontological approach has been developed and presented. The methodology describes the development of two components: a knowledge base and a combined component consisting of a solver and an explanation subsystem.