Models of recognition operators defined in the space of large dimension attributes

Abstract

The problems of constructing modified recognition operators are considered taking into account the high dimensionality of the feature space. As a basic algorithm, models of recognition algorithms based on radial functions are considered. The main idea of the model of modified recognition operators is the formation of independent subsets of interrelated objects and the allocation of basic objects for each subset of tightly coupled objects. A distinctive feature of the proposed model of algorithms is to determine the appropriate set of distance functions when building a model of discriminating operators. To test the performance of the proposed object model, experimental studies were carried out in solving a number of problems. The main advantage of the proposed recognition operators is the improvement of accuracy and a significant reduction in the volume of computational operations in recognition of unknown objects, which allows them to be used in the construction of recognition systems operating in real time.