Behaviour Based Credit Card Fraud Detection Design And Analysis By Using Deep Stacked Autoencoder Based Harris Grey Wolf (Hgw) Method

Abstract

A country's economic development and progress are greatly influenced by its banking and financial industry. Recent years have seen a dramatic growth in the number of customers making purchases with plastic, whether they do it online or in a brick-and-mortar store. Many problems are being caused by fraudsters to clients, banks, and other financial institutions. As more and more people gain access to high-speed Internet, online banking expands to become a crucial trading platform. Users' confidence and safety can be compromised by fraudulent transactions and other forms of fake banking activity. Further, with the rise of sophisticated frauds like malware infection, scams, and bogus websites, businesses lose a great deal of money to fraud. In order to detect fraudulent activity in credit card transactions, this study develops three new contributions. The first enhancement is a data-balanced Deep stacking autoencoder for fraud detection based on the Harris Grey Wolf (HGW) network. Harris Grey Wolf is proposed to train Deep stacked autoencoder in this case (HGW). The suggested HGW-based Deep stacked autoencoder provides the most effective method for uncovering frauds by applying a fitness function, adapting to produce minimal error, and computing the optimal solution over a series of iterations. Using transaction data, we select the most important features to use for detection since they increase the rate and accuracy of detection. Features were selected because of their usefulness in providing crucial data and improving detection accuracy. This is how the efficacy of fraud detection systems is validated and distinguished from genuine systems. Accuracy is 0.92, sensitivity is 0.76, and specificity is 0.92 for HGW Deep Stacked autoencoder across a range of performance parameters