Construction of an ensemble of logical correctors on the basis of elementary classifiers

Abstract

A problem of constructing correct recognition algorithms on the basis of incorrect elementary classifiers is considered. A model of recognition procedures based on the construction of a family of logical correctors is proposed and analyzed. To this end, a genetic approach is applied that allows one, first, to reduce the computational cost and, second, to construct correctors with high recognition ability. This model is tested on real problems.