As the number of users opting for credit card payment is increasing daily worldwide, the threats posed by internet fraudsters on this type of payment are also on the increase. Banks, merchants and consumers globally have lost billions of... more
As the number of users opting for credit card payment is increasing daily worldwide, the threats posed by internet fraudsters on this type of payment are also on the increase. Banks, merchants and consumers globally have lost billions of dollars as a result of this type of fraud. The shortcomings of many of the existing credit card fraud detection techniques include their inability to effectively detect fraudulent transactions, the high false alarm rate, and high computational cost. These necessitated the development of more efficient credit card fraud prevention measures. Many models have been developed in the literature; however, the accuracy of the model is critical. In this paper, fraud detection model using K-Star machine learning algorithm is presented and the performance is evaluated using German Credit and Australian Credit datasets. The algorithm proposed in this paper proved to be highly effective and efficient with a resultant classification accuracy of 100%, very low false positive rate (0.00) and very high true positive rate of 1.00. All experiments are conducted on WEKA data mining and machine learning simulation environment.
