nificant class imbalance, with fraudulent instances representing a negligible fraction relative to legitimate ones. To reduce this, resampling and algorithmic balancing techniques are employed. The Synthetic Minority Over-Sampling Technique (SMOTE) creates fake minority samples by feature space interpolation, whereas Random Under sampling (RUS) reduces majority occurrences to preserve proportionality. Cost-sensitive learning modifies model penalties to highlight fraud detection while maintaining overall accuracy. The integrated method improves classifier sensitivity