ASN-SMOTE: a synthetic minority oversampling method with adaptive qualified synthesizer selection

Abstract Oversampling is a promising preprocessing technique for imbalanced datasets which generates new minority instances to balance the dataset. However, improper generated instances, i.e., noise may interfere learning of classifier and impact it negatively. Given this, in this paper, we propose simple effective oversampling approach known as ASN-SMOTE based on k -nearest neighbors synthetic technology (SMOTE). first filters class by determining whether nearest neighbor each instance belongs or majority class. After that, uses effectively perceive decision boundary, inside qualified are selected adaptively proposed adaptive selection scheme synthesize instance. To substantiate effectiveness, has been applied three different classifiers comprehensive experiments have conducted 24 benchmark datasets. also extensively compared with nine notable algorithms. The results show that achieves best implementation available at: https://www.github.com/yixinkai123/ASN-SMOTE/ .