Atlas · GenAI 2026
SMOTE
Generating synthetic minority-class training examples from neighboring minority samples to address class imbalance, with resampling confined to training folds.
Also searchable as: Synthetic Minority Over-sampling Technique, SMOTE Oversampling
conceptModel Selection & TuningAI consensus: 0/3
Prerequisites
- mediumClassification
Class labels and imbalance define the intervention.
Recommended reference
imbalanced-learn: SMOTE — https://imbalanced-learn.org/stable/references/generated/imblearn.over_sampling.SMOTE.html; scikit-learn: Common pitfalls and recommended practices — https://scikit-learn.org/stable/common_pitfalls.html#data-leakage
Reviewed sources
Primary and first-party material reviewed for this editorial summary. These citations are separate from the AI consensus score above.
- imbalanced-learn: SMOTE
Synthetic minority over-sampling based on neighboring minority examples.
- scikit-learn: Common pitfalls and recommended practices
Training-only fitting of transformations and consistent use at evaluation and prediction time.
- Chawla et al.: SMOTE Synthetic Minority Over-sampling Technique
The original synthetic minority over-sampling method and evaluation.
Notes from AI deep research
Related skills
- → is subcategory of: Class Imbalance Handling