Kudhair A. Predicate conceptual models and method of structured objects prototyping in artificial intelligence systems

The purpose of research is development of representation predicate models and forming method of a structured object generalized prototypes in order to allow them to search and recognition taking into account the changes in the properties of these objects. Main results. First proposed the predicate model of set-theoretic concepts of primary, which allows us to describe a system of concepts by means of their properties, represented by the equations of algebra of predicate operations. The model allows generalizing the concept by establishing logical connections between their properties. An improved method of predicate prototype structured object constructing differs from the existing rules of the objects structure generalization concepts formation use, based on the concepts properties. This allows to iteratively refining the prototype due to additions of these properties. It has been further developed predicate model of logical connections between concepts. Models allow improvements representation of domain knowledge by stepping association of concepts in the group with the help of logical links. It has been further developed predicate model of knowledge representation on the prototype of a structured object, which, in contrast to the existing, describes the object as a set of concepts, logical connections between these concepts, as well as between their properties. The model allows building as a multi-level representation of an object, which reflects both the structure and the hierarchy of concepts, it is described, as well as their properties.