METAMODELS FOR SPECIFICATIONS AND CODE GENERATION OF KNOWLEDGE BASES
Nikita O. Dorodnykh, Olga A. Nikolaychuk, Alexander Yu. Yurin, Sergey A. Korshunov
Matrosov Institute for System Dynamics and Control Theory of Siberian Branch of Russian Academy of Sciences (ISDCT SB RAS), CentrSib LLC
The paper discusses the problem of improving the knowledge bases and intelligent systems design. The efficiency of this process can be improved with the aid of automatic generation of specifications and source codes of knowledge bases for a certain knowledge representation language. The generation, in turn, is based on the analysis and transformation of various information sources, such as conceptual models that describe a subject domain and are presented in different formats (e.g., UML models, concept maps, tree-like semantic structures, fishbone diagrams, etc.). Generation of specifications and codes methodologically is based on the model transformation concept from a model-driven engineering (MDE) and uses metamodels to describe the source and target (CLIPS and OWL) formalisms. The paper describes a method for computer-aided metamodel formation for the analyzed formats of conceptual models, and meta-metamodels for the unified representation and storing of metamodels.
intelligent system, knowledge base, model-driven engineering, metamodeling, model transformation, metamodel, conceptual model, code generation