Credit Card Fraud Prediction System Using Machine Learning Algorithms

Abstract

In today’s world, credit card is found to be one of the most commonly used system for online transactions due to its convenience of usage. Howe ver, this ease of usage comes with its own share of troubles. Recent statistics have found that global credit card fraud

losses equaled to be $1.48 billion in 2018. Increasing participation in online transactions raises fraudulent cases globally. This causes tremendous losses to users, banks and merchants. These frauds can be detected, if adequate amount of data is collected and preprocessed and fed into machine learning algorithms. This project aims to apply different Machine Learning algorithms (Logistic Regression, Random Forest and Naïve Bayes) to find the most efficient algorithm in detecting the fraudulent transactions based on selected parameters. And this algorithm is used to predict the probability of a transaction being fraudulent. This not only secures the users’ money but also allows the merchants to have fraud-free business and the banks to work smoothly to help the users. The significance of the project is that it brings into light the fraud that is going on in the real world to earn money illegally. It secures the digital life of a person in terms of money and hence, detects the credit card misuse.