Credit Card Fraud Detection Using Random Forest and Cart Algorithm

Abstract

The project is primarily concerned with detecting credit card fraud in the real world. The phenomenal increase in the number of credit card transactions has recently resulted in a significant increase in fraudulent activities. The goal is to obtain goods without paying for them or to withdraw unauthorized funds from an account. In order to minimize losses, all credit card issuing banks must implement effective fraud detection systems. One of the most difficult challenges in starting the business is that neither the card nor the cardholder must be present at the time of purchase. This makes it impossible for the merchant to determine whether or not the customer making a purchase is the legitimate cardholder. Using the proposed scheme and the random forest algorithm, the accuracy of detecting fraud can be improved. Random forest algorithm classification process to analyze data set and user current dataset. Finally, improve the accuracy of the output data. The techniques' performance is measured using accuracy, sensitivity, specificity, and precision. Then, by processing some of the provided attributes, the fraud detection is identified and the graphical model visualization is provided. The techniques' performance is measured using accuracy, sensitivity, specificity, and precision.