How Fraud Detection Systems Use Graph Databases

Fraud is a billion-dollar business and it is increasing every year. The PwC global economic crime survey of 2009 suggests that close to 30 percent of companies worldwide have reported being victims of fraud in the past years. Since we are living in the digitalized world where technology evolves fast and number of transactions grows rapidly, the development of new technologies has provided fraudsters more ways to smuggle money out from person’s or organization’s pocket via using IT system vulnerability or snatching digitalized transactions. Fraud Detection System is developed to detect and prevent such attempts from fraudsters with features like analyzing and monitoring malice transactions.

Early fraud detecting techniques were oriented toward extracting quantitive and statistical data characteristics. These techniques facilitate useful data interpretations and can help to get better insights into the processes behind the data to detec and prevent frauds. However, it required numbers of human analysts and was a time consuming task. Therefore, many fraud detection systems are using statistical techniques combining with machine learning algorithms nowadays. They detects frauds by collecting data, calculating various statical parameters such as averages, quantiles, probability and etc., separating malice transactions from normal transactions. Compare to the traditional fraud detection techniques, they facilitate automated detection and preventing of fraud activities and, more importantly, could provide such services in real-time.

It is true that conventional fraud detection system did a quiet good job to protect organizations from fraud attempts. However, since fraud involves one or more persons who intentionally act secretly to deprive another of something of value, for their own benefit, conventional way of calculating probability and discrete analyses is becoming hard to prevent advanced fraud techniques, which is to exploit the weaknesses of discrete analysis. Moreover, since speed of network and computer are growing faster, organizations have less time to analyze fast passing transactions, and it makes harder to catch fraud activities in real-time. To solve above mentioned problems and enhance their systems, an increasing number of companies are starting to use graph databases.Unlike conventional ways of looking at data, graphs are designed to express relatedness. Graph databases uncover patterns that are difficult to detect using conventional techniques such as relational databases.

While most of fraud activities can and does involve criminal rings, even a single well-informed fraudster can create a large number of synthetic identities and to carry out sizeable schemes. Consider an online transaction with the following identifiers: user ID, IP address, geo location, a tracking cookie and a credit card number. Typically, the relationships between these identifiers should be one-to-one(well, most of the time). Some variations naturally account for shared machines, families sharing a single credit card number, individuals using multiple computers and such. However, as soon as the relationships between these variable exceed a reasonable number, fraud should be considered as a strong possibility. The more interconnections exist amongst identifiers, the greater the cause for concern. Large and tightly-knit graphs are very indicators that fraud is taking place. Thus, by putting checks and event triggers into pathways of transactions, we can detect, monitor possible fraud activities and catch fraudsters’ malice attempts on people and organizations in real time.