Graph technology discovers anomalies hidden throughout the network of information.
Fraud detection solution detects and prevents suspicious behaviors found in fraud rings.
Fraud Detection®
CASE STUDIES
Illegal Activity Detection System
Many smugglers and their accomplices avoid the radar of the K Agency by hiding among the flight travelers while illegally smuggling gold and drugs. A graph-based fraud detection system can monitor and quickly analyze relationships from a vast network of data. The information analysts sought an analysis system that could assist them in monitoring and pinpointing the movement of these smuggling offenders.
Cyber Threat Response System
Bitnine’s graph-based CTI system provides techniques for analyzing cyber attacks and identifying attackers based on graph models and algorithms. The CTI platform share reliable cyber threat information in a standardized format, visualizes cyber attack patterns in real-time, and recommends an appropriate countermeasure based on the importance of information.
Exposing Fake Deposit with Transaction Pattern Analysis
Bitnine’s AgensGraph was used in a certain bank’s AI-based voice phishing monitoring system. The AI and big data reliant system was built to deal with ever-growing fraud methods. As a result of the pilot operation for two months, the number of fake deposit occurrences decreased by about 42% compared to the period of operation with the existing system.
FDS with Graph Modeling
Written as an introduction to the white paper on Graph FDS, this use case reviews how graph modeling can be applied in the FDS scenario for detailed analysis. Expressing transaction patterns in nodes and edges would not only be intuitive, but also narrow down suspicious activities previously undetected in the form of legacy FDS.
① G-FDS: Insurance Scam
The first scenario of the white paper on Graph FDS shows how graph modeling can expose potential insurance scams. The scenario intends to show how an assailant and victim of the car accident are actually partners in crime and how they purposefully caused similar car accidents to get insurance fees. The graph modeling is used to prove the relationships between these people, show their past records of similar accidents, and expose third person that may be related to their deeds, etc.
② G-FDS: Credit Card Fraud
The second scenario of the white paper on Graph FDS shows how graph modeling can expose credit card fraud. The scenario sets a certain rule that would be considered fraudulent activity. For example, if a supposed counterfeit credit card is used to purchase a luxury brand-name product, the card company is likely to contact the card owner about the large payment made. So the rules are set in which the fraudsters would divide the purchases by buying multiple of the same items with different counterfeit cards. If such transactions happen and match the previously set rule, a graph modeling is used to double-check whether the credit card transactions are in fact, made with counterfeit cards.
③ G-FDS: Illegal Drawback
The third scenario of the white paper on Graph FDS shows how graph modeling can expose illegal drawbacks. In this scenario, a paper company or representative with a criminal record, etc is assumed to issue an illegal drawback. In order to prove these companies’ activities are illegal, all their payment history are tracked and monitored via graph modeling for intuitive understanding.
④ G-FDS: Bank Account Theft
The fourth and final scenario of the white paper on Graph FDS shows how graph modeling can expose bank account theft. Graph modeling is used to express suspicious transactions and detect fake deposits, burner phones, bank accounts, etc. The fraud methods are getting advanced day by day and graph FDS helps to catch up to the intelligence of modern fraudsters.
