DESIGN AND DEVELOPMENT OF AN ENHANCED MODEL FOR CREDIT CARD FRAUD DETECTION IN BANKS

SOURCE:

Faculty: Physical Sciences
Department: Computer Science

CONTRIBUTORS:

Amanze B. Chibuike
Inyiama H.C.

ABSTRACT:

The growth in electronic transactions has resulted in a greater demand for fast and accurate user identification and authentication. Conventional method of identification based on possession of pin and password are not all together reliable. Higher acceptability and convenience of credit card for purchases have not only given personal comfort to customers but also attracted a large number of attackers. As a result, credit card payment systems must be supported by efficient fraud detection capability for minimizing unwanted activities by fraudsters. Most of the well-known algorithms for fraud detection are based on supervised training. Every cardholder has a certain shopping behaviour, which establishes an activity profile for him. Existing fraud detection systems try to capture behavioural patterns as rules which are static. This becomes ineffective when cardholder develops new patterns. This dissertation aimed at designing and developing an enhanced model for credit card fraud detection in Nigerian banks using Machine Learning Techniques and multi- agents that combine evidence from current as well as past behaviour to determine the suspicions level of each incoming transaction. The model was designed using Object-Oriented Analysis and Design Methodology (OOADM), Multi-Agent Methodology and Machine Learning Technique respectively. The model was programmed and implemented using PHP while the database was implemented with MySQL. Test results on the new system using confusion matrix shows a significant positive impact 94% accuracy in credit card fraud detection as against 57% of accuracy by the existing system, and hence a significant improvement on overall operating efficiency. Thus, the new credit card fraud (CCF) detection system using multi-agents is compatible with other detection software but has significantly higher performance efficiency (94%). The model is therefore recommended for use by banks, financial agencies and government agencies.