Fraud Detection – A Hybrid Machine Learning Approach

Online Fraud is an often a major crime in online money transactions between a sender and recipient. While the transaction occurs online, some incidents may occur where the transactions are made without sender’s consent. There is a potential of many of us who becomes victim or may can be which is intangible and dynamic. This paper argues and help with the methods to prevent from happening the fraudulent. Once the fraud occurs there is less chances for someone to get their hard-earned money back, whereas some of them had also lost their lifetime savings. However, the decisions over it should be more transparent and accurate to win the trust of regulators, businesses and bank. Also, it is a critical issue in financial transactions, especially with the rise of digital payments. The data for this study is obtained with the help of real-time transactions which shows daily transactions scheme known as PaySim. It is a simulator for mobile money transactions, provides a realistic dataset to analyse fraud patterns. At the same time fraud detection methods will help to resolve the issue. Whereas the notorious fraud detection dataset is well known for its several legitimate transactions rather than fraudulent ones. Our paper aims to use the methods of machine learning which helps to predict fraud transactions occurs. Although the Online Fraudulent of money is crime in world but criminals often do this without a guilt which comes under IT ACT, 2000 and PREVENTION OF MONEY LAUNDERING ACT(PMLA) 2000.