Mini toolsDefinition
A technique for handling imbalanced datasets by generating synthetic examples of the minority class.
Detailed Explanation
SMOTE (Synthetic Minority Over-sampling Technique) works by selecting examples that are close in the feature space drawing a line between the examples in the feature space and drawing a new sample at a point along that line. This creates synthetic examples rather than simply duplicating existing ones helping to avoid overfitting.
Use Cases
Biomedical classification fraud detection rare event prediction
