International Journal of Security, Privacy and Trust Management ( IJSPTM), Vol. 1, No 5, October 2012 DOI : 10.5121/ijsptm.2012.1501 1         Ali Ahmadian Ramaki 1 , Reza Asgari 2 and Reza Ebrahimi Atani 3 1 Department of Computer Engineering, Guilan University, Rasht, Iran ahmadianrali@msc.guilan.ac.ir 2 Department of Computer Engineering, Guilan University, Rasht, Iran rezaasgari@msc.guilan.ac.ir 3 Department of Computer Engineering, Guilan University, Rasht, Iran rebrahimi@guilan.ac.ir ABSTRACT Using graphs as to extracting and presenting data has a wide range of applications. Such applications may appear in detecting semantic and structural patters and exploiting graphs toward such applications have steadily been growing. In this paper we are going to display one of the most perilous abnormalities in credit cards industry on such concept basis. With advancing technology in field of banking, the rate of use of credit cards has remarkably been escalated. Correspondingly frauds frequency have increased in this area which to surmount such anomalies we model them by means of graphs. Of the prominent advantage of proposed approach is drop of system overload rate during running computations in order to detecting frauds and consequently acceleration of detection speed. KEYWORDS Credit Card, Fraud, Legal Transaction, Fraudulent Transaction, Graph Model 1. INTRODUCTION Today one of the biggest threats to commercial institutes is fraud in credit cards. Understanding of fraud mechanism for fighting back its effects is subsequently a necessarily salient task. Fraudsters swindling by credit cards take advantage of disperse methods to perpetrate such illegal acts [1]. Not long ago banking researches have been performed toward morality in banking industry and proportionately fraudulent acts have become intricate. Fraud in fact connotes gaining a service, commodity and money by means of dishonest methods growing all over the world. Fraud as a crime is occasionally hard to detect. Such fraud may occur through any type of credit production like personal loans, home loans or retailing. Moreover, methods to fraud have marvellously been expanded in parallel to technology promotion. A necessity hence, for different prominent businesses such as banking is to applying systems and processes to thwart fraud in their business field. Varied methods to fraud detection have been developed each of which has its own pros and cons [2]. As to significance of transactions performance in such field alongside with synchronization of system responses against every transaction performed by credit cards owners different approaches are practiced. Of the most important approaches we can list data mining and neural networks [3]. Their most remarkable disadvantages are high computational overload and time-consuming detection process