Credit Card Fraud Mitigation System

Abstract

Consumers rely on the Internet to shop, bank and invest online. Credit-card transactions are to be the most critical transaction in the world. We use it for meals when dining out, pay for our gas at the pump, and purchases. But we never think, or even care, about the process running in the background. We give our cards to supermarket cashiers, Petrol pumpers and waiters. And they have a chance of copying card information and CVC number on it. If it s through Internet transaction there are many ways of tapping and stealing that information. Here my role is to use customer pattern identification, build a signature using data mining techniques. Different oracle functions used to do the calculations. I selected this method because of easy implementation. Our credit-card database is implemented on Oracle. For the Developments I used Bayesian learning and Decision trees techniques. Bayesian learning in which evidences from current as well as past behavior are combined together and depending on certain type of shopping behavior establishes an activity profile for every cardholder. Decision trees are statistical data mining technique that express independent attributes and a dependent attributes logically AND in a tree shaped structure. Classification rules, extracted from decision trees, are IF-THEN expressions and all the tests have to succeed if each rule is to be generated. In this system current transaction will be compared with customer signature record. If some unusual transaction occurs it will notify the systems operator or to the person who is in charge. For data mining I used WEKA installed on Linux as experimental work and Octave for neural networks.