A Comparison Study of Fraud Detection in Usage of Credit Cards using Machine Learning

Abstract

Credit Cards are widely used in online transactions due to its usefulness and effectiveness. Despite the hype, the credit card fraudulent activities have inconsistent patterns of behaviour and frequently change them. Fraudsters are aware of new technologies to hack the online transactions. Credit Card Fraud is becoming more common due to the increasing use of credit cards. Designing and Creating a cutting-edge Machine Learning based Fraud Detection solution is the major goal of this research work. Machine learning forms the basis of the algorithms like XGBoost, Support Vector Machine, Decision Tree, Random Forest, and Logistic Regression. A CNN (Convolutional Neural Network) is then used to enhance the efficiency of fraud detection. Layers of a Convolutional Neural Network assists in obtaining accurate detection. An extensive empirical analysis was conducted using the most recent CNN model's hidden layer counts, epochs, and applications. The F1 score, accuracy, precision, and recall all affect the results of the algorithms. Area Under Curves (AUC) has been altered to leverage values of 99.9%, 85.71%, 93% and 98%. A ROC curve is produced using the confusion matrix as a framework. The proposed method overcomes the problem of Credit Card identification by combining Deep Learning with Machine Learning techniques. Furthermore, to reduce the number of false negatives, this study has performed data-matching trials with the implementation of Deep Learning Techniques. This study has employed Deep Learning techniques and perform trials to match the data with the model. Utilizing the suggested strategy, it is possible to locate Credit Card Fraud (CCF) remotely from any location.