Application of Computational Intelligence to Identify Credit Card Fraud

Recently, computational intelligence algorithms have broadly used in many intelligence and expert systems. This paper tests the usefulness of computational intelligence for detecting credit card anomaly. UCSD2009, UC San Diego data mining contest's datasets were used to detect credit card anomaly. Different classification algorithms of computational intelligence were applied on the dataset to measure the performance. The experimental result found that these classifiers perform better to detect fraud but the misclassification rate is still so high. Among all the classifiers, MultiClassClassifier performs better because this classifier detects most of the anomalies.