ADVANCED CREDIT CARD FRAUD DETECTION USING ODDITY DISCOVERY TECHNIQUES

ADVANCED CREDIT CARD FRAUD DETECTION USING ODDITY DISCOVERY TECHNIQUES

Muthamilselvi S

Computer Science and

Engineering
E.G.S. Pillay Engineering College
Nagapattinam, India
tamizhselvakumar003@gmail.com

Surya M

Computer Science and

Engineering
E.G.S. Pillay Engineering College
Nagapattinam, India
suryamuruga500@gmail.com

Dr.T.Ganesan,M.E, Ph.D.,

Computer Science and

Engineering
E.G.S. Pillay Engineering College
Nagapattinam, India
ganesan.t@egspec.org

Abstract—Credit cards are the most widely accepted payment method for both disconnected and online transactions. In the banking industry's fight against card crime, automatic systems to detect and prevent card fraud are an important weapon. Since machine learning and Oddity Discovery Techniques (ODT) are efficient technologies and methodologies that are also simple to implement, they are used in the field of credit card fraud detection. It is applied to lessen fraud-related behaviors. The major goal of this project was to develop software that can recognize potentially fraudulent credit card transactions in a given data set and evaluate its performance against other classifiers using evaluation metrics. The given data set was used to train the software, which was built using some of the most well-liked machine learning and deep learning classification methods, including random forests, isolation forests, and neural networks. The Python programming language was used to implement the project, and load balancing and feature selection were maintained throughout. There was, however, no autonomous system that could definitively classify a transaction as fraudulent. The goal was to draw attention to transactions that, according to some known or otherwise learned criteria, have a high likelihood of being fraudulent.

Keywords: Credit card, Classifiers, ML- Machine Learning, Fraud detection, Feature Selection.