Fig. 1
From: An approach for handling imbalanced datasets using borderline shifting
Illustration of the Synthetic Minority Over-sampling Technique (SMOTE). The left panel shows the original imbalanced dataset, where minority class samples (red triangles) are underrepresented relative to the majority class (blue circles). The middle panel depicts the SMOTE process, in which a minority instance is selected and synthetic samples are generated along the line segments joining its nearest minority neighbors. The right panel presents the resulting balanced dataset after oversampling, demonstrating how SMOTE increases minority class density to reduce class imbalance27 and quick technique, it may lead to the loss of important information. As discussed by28, its effectiveness depends on maintaining representative samples from the majority class during reduction.
