Intelligent Systems Based on Ontology Representation Transformations

The growing complexity of information systems requires offloading knowledge-based decision making to software. The common approach is to encode formalized domain knowledge in form of ontology. However, this approach has many inherent flaws, such as subjectivity of ontology design, difficulties to formalize different perspectives and points of view for the same domain, complexity and large size of ontologies. This problem can be resolved by using multiple networked ontologies. The ultimate criterion for successful ontology design is provided by the effectiveness and efficiency of resolving real life problems using ontology. On the other hand, the efficiency of ontology usage depends also on language used to formulate ontology. Each language is geared towards the specific application area and has limitations which make it a poor choice for problems in another area. Therefore, the versatile knowledge based system should be able change knowledge representation form on task by task basis. The structure and functions of such a system are described. For the central repository of knowledge it is proposed to use a language, based on multisort algebras. In order to represent language transformations this algebraic language is enhanced by adding mappings domain. As an example, the mapping between RDF schema and algebraic language is provided.