Oversampling Approach Using Radius-SMOTE for Imbalance Electroencephalography Datasets

Several studies related to emotion recognition based on Electroencephalogram signals have been carried out in feature extraction, representation, and classification. However, is strongly influenced by the distribution or balance of data. On other hand, limited data obtained significantly affects imbalance condition resulting signal It has an impact low accuracy recognition. Therefore, these problems, contribution this research propose Radius SMOTE method overcome DEAP dataset process. In addition EEG oversampling process, there are several vital processes signals, including extraction process classification This study uses Differential Entropy (DE) The compares two methods, namely Decision Tree Convolutional Neural Network method. Based using method, application with resulted recognizing arousal valence emotions 78.78% 75.14%, respectively. Meanwhile, can accurately identify 82.10% 78.99%, Doi: 10.28991/ESJ-2022-06-02-013 Full Text: PDF