SMOTE (Synthetic Minority Over-sampling Technique) is a popular method used to address class imbalance in datasets by generating synthetic samples for the minority class. Variants of SMOTE have been developed to enhance its effectiveness, including Borderline-SMOTE, which focuses on generating samples near the decision boundary, and ADASYN, which adapts the number of synthetic samples based on the density of minority class instances. Other notable variants include KMeans-SMOTE, which combines clustering with SMOTE to create samples based on cluster centroids, and SMOTE-ENC, designed specifically for datasets with categorical features. These variants aim to improve model performance by providing more representative training data.