Imbalanced Classification Strategy BuilderPREMIUM

## CONTEXT
Fraud, churn, disease, and defect detection share a brutal trait: the positive class is rare, and naive models achieve 99% accuracy by predicting "no" every time. In 2026 the standard mistakes persist: optimizing accuracy, blindly applying SMOTE, and ignoring the gap between resampled training distributions…