Performance Evaluation Of Hybrid Machine Learning Models For Credit Card Fraud Detection

Abstract

Credit Card Fraud Cares With The Illegal Use Of Master Card Information For Purchases. Credit Card Transactions Are Often Accomplished Either Physically Or Digitally. In The Manual Transactions, The Credit Card Is Included During The Transactions. In Digital Transactions, This Will Happen Over The Phone Or The Web. Cardholders Might Be Providing Their Card Number, Expiry Date, And The Verification Of The Card Number Through Telephone Or Website. Billions Of Dollars Are Lost Thanks To Master Card Fraud Per Annum. Machine Learning Techniques Are Wont To Detect Master Card Fraud. Standard Models Are First Used. Then, Hybrid Methods Which Use Random Forest And Xgboost Segmentation And Popular Voting Method Are Applied. Then, A Real-world Master Card Data Set From A Financial Organization Is Analysed. In Addition, Noise Is Added To The Info Samples To Further Assess The Robustness Of The Algorithms. Here Random Forest Segmentation And Xgboost Algorithm Will Give The 94% Percent Accuracy.

Keywords