SVM classification: Optimization with the SMOTE algorithm for the class imbalance problem
Liliya Anatolievna Demidova, Irina Klyueva · 2017
In this paper a new approach to selection of the optimal parameters values for the SMOTE (Synthetic Minority Over-sampling Technique) algorithm in the problem of the SVM (Support Vector Machine) classification of imbalanced datasets has been suggested. This approach allows reducing the time expenditures for the search of the optimum parameters values of the SMOTE algorithm. The experimental results show that the offered approach allows increasing the classification quality of the SVM classifier.