An Intelligent Method for Credit Card Fraud Detection using Improved CNN and Extreme Learning Machine

Abstract

Credit card fraud has been on the rise in recent years. Criminals are taking advantage of the public by pretending to be legitimate businesses or individuals in order to steal their money. So, it is essential to combat this type of fraud. This research study has developed a novel method for spotting suspicious credit card transactions. The proposed approach may provide most of the necessary information to spot illegal or fraudulent financial transactions. As technology advances at an unprecedented rate, it becomes more difficult to monitor the illegal transactions and money transfers. With the recent developments in machine learning, artificial intelligence, and other critical areas of IT, it is now possible to automate this process and save huge amounts of labor involve in recognizing credit card fraud. After receiving an input image, it is preprocessed and features are retrieved by using principal component analysis, and then the data is used for training the CNN-ELM model. The proposed method achieves a higher accuracy (about 98.7%) than other methods like CNN and ELM.