How do you handle imbalanced datasets in classification problems?

Imbalanced datasets occur when class distribution is skewed, leading to biased models that favor majority classes.\n\n• Resampling: SMOTE for oversampling, random undersampling\n• Algorithm-level: Class weights, cost-sensitive learning\n• Evaluation: Use F1-score, precision-recall curves, not just accuracy\n• Ensemble: Balanced bagging, EasyEnsemble\n\nExample: Fraud detection with 1% positive cases. Apply SMOTE to generate synthetic minority samples, use class_weight="balanced" in sklearn, and evaluate with ROC-AUC and precision-recall metrics.