ARTIFICIAL NEURAL NETWORK APPROACH FOR DETECTION OF CREDIT CARD FRAUD USING CUSTOMER BEHAVIOUR.

Abstract

Credit cards serve as a good substitute for cash because credit cards are made of plastic and it is distributed by financial institutions like banks to make transactions simple and convenient without having to carry cash. Changes in technology, particularly the internet, have made credit card use more prevalent as well as exposed Credit card to unauthorized users. The abuse by fraudsters can be checked by taking the required preventative steps, and it is possible to research the behavior of such fraudulent operations in order to minimize them and prevent recurrence. The user's behavior and location are scanned as part of the credit card fraud detection so as to look for odd patterns. The patterns contain user traits like spending habits and geographic areas to confirm the user's identification. The system needs to be re verified if any odd patterns are discovered, since people typically display particular behavioral patterns. Old in pattern discovered based on past purchase data of cardholders is analyzed. A set of patterns containing data on the typical purchase and the amount paid can be used to represent each cardholder. This system was designed on artificial neural network (ANN) algorithm, which provides better accuracy close to 100%. It offers greater accuracy compared to unsupervised learning algorithms.