This email address is being protected from spambots. You need JavaScript enabled to view it.
 
+7 (4912) 72-03-73
 
Интернет-портал РГРТУ: http://rsreu.ru

UDC 004.855.5

RANDOM FOREST ALGORITHM IN THE PROBLEM OF IMPROVING SVM-CLASSIFICATION QUALITY

L. A. Demidova, PhD (technical sciences), full professor, RSREU, Ryazan; This email address is being protected from spambots. You need JavaScript enabled to view it.
I. А. Klyueva, post-graduate student, RSREU, Ryazan; This email address is being protected from spambots. You need JavaScript enabled to view it..

The aim of the work is to improve the quality of objects SVM-classification (Support Vector Machine) by hybridizing the SVM-classifier with the random forest classifier (Random Forest, RF) used as an auxiliary. Clarification of the classification solutions obtained during the development of the SVM-classifier on the basis of initial dataset is performed for the objects located in experimentally determined subareas near the hyperplane separating the classes and including both correctly and erroneously classified objects. In the case of improving the quality of objects classification from the initial dataset, the proposed hybrid approach to the classification of objects is recommended for the classification of new objects. When developing the SVM-classifier, fixed default parameter values are used. A comparative analysis of classification results obtained during computational experiments in the hybridization of the SVM-classifier with two auxiliary classifiers – random forest classifier (RF-classifier) and k-nearest neighbor classifier (kNN-classifier), for which the parameter values are determined randomly, confirms the expediency of using these classifiers to increase quality SVM-classification. It was found that in most cases, the random forest classifier works better in terms of improving the quality of SVM-classification in comparison with kNN-classifier.

Key words: intelligent classification algorithm, SVM-algorithm, k-nearest neighbors algorithm, random forest algorithm, support vector, kernel function, kernel function parameter, regularization parameter, hybridization.

 Download