Machine Learning based Credit Card Fraud Detection - A Review

The current era is known as the era of digital technology, where everything around us is shifting its functioning towards digital world, be it a space mission or a tea seller selling his tea, everyone is linked with the digital world in one way or other. In the same manner, the way of transaction of money is also shifting from physical currency to digital currency which has a lot of benefits associated with it like, saving the material and manufacturing cost, manpower etc. however everything in this world comes with both pros and cons and hence the frauds are also evolving with time. One of such frauds which is causing huge disturbance in world's economy is credit card frauds that takes place frequently resulting in a large amount of financial loss. The quantity of transactions occurring online on day-to-day basis is increasing exponentially. This review paper primarily throws light upon the four main types of fraud instances in present-time transactions. Every fraud mentioned is tackled using sequence of machine learning models and the most appropriate methods are selected via an evaluation. To make it an actual time fraud detection, the aid of predicting analysis carried out by implementing API module and machine learning models are taken to decide whether a transaction is fraudulent or legitimate. A suitable strategy is used that efficiently addresses the undistributed collection of data in this review paper.