Abstract :

Credit Card Fraud Is A Big Problem Now That More And More People Are Buying Things Online. It Costs A Lot Of Money For Both Individuals And Businesses. It Is Hard To Find Fraud Since The Datasets Are Very Unbalanced, With Only A Small Number Of Illicit Transactions Compared To Genuine Ones. Fixing This Mismatch Is Very Important For Making Fraud Detection Systems That Work Well And Give Accurate Results. To Solve This Problem, This Study Looks Into Advanced Hybrid Undersampling And Oversampling Methods That Can Help Find Fake Transactions While Still Keeping Excellent Performance Across A Range Of Assessment Criteria. To Balance The Dataset And Make It Easier To Work With, We Used Sampling Methods Including SMOTE, B-SMOTE, ADASYN, SMOTE-Tomek, SMOTEEEN, And Hybrid BIRCH Borderline SMOTE. The Voting Classifier, Which Combines Boosted Decision Trees And ExtraTree, Always Did Better Than The Other Models In Terms Of Accuracy, Precision, Recall, And F1-score Across All Sampling Methods. This Shows That It Is Strong And Works Well With Credit Card Fraud Datasets That Are Not Balanced. The Results Show That Using A Mix Of Hybrid Sampling Approaches And Ensemble Learning Could Make Fraud Detection Systems A Lot Better. Index Terms - Borderline SMOTE, Class Imbalance, Credit Card, Fraud Detection, Sampling Techniques, Tomek Links

Published:

22-7-2025

Issue:

Vol. 25 No. 7 (2025)

Page Nos:

751 - 763

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite