Improvement the Imbalanced Data Classification with Restarting Genetic Algorithm for Support Vector Machine Algorithm

Abstract

The general datamining algorithm also classify the balanced dataset, when the data have imbalanced the predicted rate over minority class is still low. The random sampling techniques has been applying to solve the imbalanced data, but sometimes the random technique has selected the features is clearly different from both, when the unseen data (from minority class) has features look like the majority class, the classification model show miss classification because the model learning sample data does not complete. To improve the performance to classify the data, the genetic algorithm is applying to finding the optimal parameter, but sometimes the genetic algorithm cannot find the best set of parameters because the random initial population is not cover the best set of parameters, in this research proposed the techniques to guarantee the genetic algorithm can find the optimal parameter by using restarting technique to recreate the initial population when the new generation show powerful less than the old population. The results show that proposed technique can improve the performance to classify the minority class from imbalanced dataset more than the other techniques.

Multilayer Perceptron Optimization on Imbalanced Data Using SVM-SMOTE and One-Hot Encoding for Credit Card Default Prediction

“…The results of high accuracy but have a low AUC score indicate that the model is experiencing bias-to-majority when the used dataset is imbalanced and handled by the classification algorithm without any proper optimization phases (Liu et al, 2019). To cope with these problems, random sampling using synthetic minority oversampling technique (SMOTE) can be used to deal with imbalanced data (Suksut et al, 2019). The SMOTE technique has several development variants, one of them is SVM-SMOTE (Zhang et al, 2019).…”