Building the rule knowledge bases on the basis of UML models.
Dorodnykh N.O., Yurin A.Yu.
The paper considers software (Personal Knowledge Base Designer, PKBD) designed for development of rule knowledge bases. The C language production system (CLIPS) was selected as the targeted knowledge representation language. The software is integrated with CASE-tools (IBM Rational Rose and StarUML) in terms of UML-models import and code generation of knowledge bases in the CLIPS format. The special generalized rule-oriented model is used for the intermediate representation and storage of extracted logical rules. This model abstracts the specific features of knowledge representation languages. The software is focused on the non-programming specialists and has an extensible architecture that implemented the ability to connect the program modules (dynamic-link libraries) supporting the code generation in other knowledge representation languages. Testing developed knowledge bases is carried out by connecting and running the rule engines. The preliminary results of software approbation for solving a variety of learning tasks are presented. Keywords: Software, knowledge bases, rules, conceptual model, code generation, CLIPS, UML.
