An Improved Oversampling Algorithm Based on the Samples’ Selection Strategy for Classifying Imbalanced Data

Figure 4

The original SVM classification effect without data processing.