Credit Card Fraud Detection Using Artificial Neural Networks and Random Forest Algorithms

Abstract

Credit Cards are one of the most well-built payment type which allows the customers to make the transactions easily and increases the purchasing power. It also offers various advantages or profits like cashback; reward points and it is very easy to use. But the main concern is Credit Card Fraud, it is increasing day by day rapidly. Credit card fraud continues to be the most popular form of identity theft worldwide. Now-a-days it has become too risky to use Credit Cards. Actually there are various types of Credit Card Frauds like Phishing and Vishing, Key Stroke Logging, POS Fraud, Application Fraud, Loss of Card or Theft. However safety measures and precautions should be taken by the individual in order to protect his/her money. It is a big business and there are organized crime rings that are behind the vast majority of this fraud their operations are industrialized, they're automated. So how to detect these frauds ? The answer is by using Machine Learning Algorithms. There is also an other way Conventional Fraud Detection, but Machine Learning Algorithms are far more accurate and precise. The main goal is to build a model which predicts whether a Transaction is fraud or not. In this project, several predictive models like Artificial Neural Networks, Random Forests, Support Vector Machine, K-Neighbors, Decision Tree, Gaussian Naive Bayes and Logistic Regression are used. The results of all these models are compared based on accuracy and the superior one is determined.