DEVELOPMENT OF THE SVM-CLASSIFIERS' ENSEMBLE USING DECORRELATION MAXIMIZATION ALGORITHM
The article deals with the problem of increasing classification accuracy with use of the SVM classifiers' ensemble. Some rules were established here to select private SVM-classifiers, which are used for creating the ensemble, as well as strategies for ensemble members’ integration.
Keywords: data classification, classification accuracy, SVM-classifier, principle of maximum decorrelation, coefficient of determination, strategy