Credit Card Fraud Detection Technique by Applying Graph Database Model

Digital transactions using credit cards are observed to be increasing day by day because of the convenience in operation. It is a matter of great concern for credit card users as well as financial institutions, providing credit card facilities for making the transactions free from possible frauds being carried out by fraudsters. The fraudsters apply different methodologies and alter their behaviours to undertake the fraudulent activities in both online and offline mode with some advanced techniques. Hence, developing a fraud detection system to identify the fraudulent activities is an important area of research to improve the credibility of credit card-based digital transactions. In this study, a fraud detection system has been proposed based on application of graph database model. The graph features being extracted using Neo4j tool are incorporated with several other features of transaction database. Subsequently, five supervised and two unsupervised machine learning algorithms are applied to them in order to detect fraudulent transactions explicitly. The features directly obtained from the transactional data are also tested with the classification models for detecting the fraudulent transactions. Critical assessment for performance of the machine learning algorithms has been carried out based on the features extracted from graph database and features extracted directly from the transaction database.