Credit Card Fraud Detection

Main Article Content

Mohit Tiwari, Vipul Sharma, Devashish Bala, Devansh, Dishant Kaushal

Abstract

The increase in the popularity of the Internet has brought about a rise in the usage of credit cards. People have shifted to e-banking due to the Covid pandemic. The rise is good as we move towards a digital India, but it has brought with it credit card fraud. There is, therefore, a requirement for a system that can detect these frauds. In this paper, we have analysed the recent work that has been done to identify credit card fraud.  We have then developed Logistic Regression and XGBoost models to detect these frauds. To make the model more efficient and robust, we have used RandomizedSearchCV to find its optimal hyper-parameters and Synthetic Minority Oversampling Technique (SMOTE) to handle the imbalanced dataset. The models were then tested on the Kaggle dataset containing over 284,000 transactions. Both achieved a very high accuracy with a ROC-AUC score of 0.99

Article Details

Section
Articles