Online transaction fraud detection techniques: A review of data mining approaches

In last decade there is a rapid advancement in e-commerce and online banking, the use of online transaction has increased. As online transaction become more popular the frauds associated with this are also rising which affects a lot to the financial industry. To overcome these problems numerous fraud detection techniques and algorithms have been proposed, data mining is used by many firms associated with fraud detection. But the data mining alone is not sufficient for detecting the fraud as it depends upon the data set containing past history of customer's transaction. This paper presents review of various fraud detection techniques and discuss the issues regarding financial dataset used in fraud detection technique, which affects the accuracy of the fraud detection and also propose a hybrid approach that uses data mining algorithm at more than one stages which contains the database level as well as network level that helps to improve the accuracy of fraud detection.