Handling Class Imbalance in Credit Card Fraud Using Various Sampling Techniques

Over the last few decades, credit card fraud (CCF) has been a severe problem for both cardholders and providers. Credit transactions are fast expanding as internet technology advances, significantly relying on internet. With advanced increased usage, rates becoming economy. However, dataset is highly imbalanced skewed. Many classification techniques used to classify non-fraud but in certain condition, they may not generate best results. Different types of sampling such under-over sampling, Synthetic Minority Oversampling, Adaptive synthetic have overcome class imbalance dataset. Then, sampled datasets classified using different machine learning like Decision Tree, Random Forest, K-Nearest Neighbors, Logistic Regression, Naive Bayes. Recall, F1- score, accuracy, precision, error rate evaluate model performance. The Regression achieved highest result with 99.94% after under Forest 99.964% over techniques.