Fraud Detection System for Payment Processing

Engineering partner:DataArt
Location:Global
Industry:Finance
Services:AI and ML

50%+

Reduction in false positives
8% of transactions flagged via anomaly detection

0.84 F-Score

Achieved on known fraud data — exceeding industry benchmarks for detection accuracy

Client

Our client provides a comprehensive suite of payment processing services. To enhance transaction security and protect customer data and transactions more effectively, it aimed to modernize its fraud detection capabilities.

Challenge

The client previously relied on a third-party fraud detection system that required sharing sensitive transaction data with external parties. However, this approach was not acceptable due to security and privacy concerns. To overcome this, the client required an in-house fraud detection system that would allow to retain full control over their data.

Solution

DataArt developed a custom, on-premises fraud detection system, ensuring all sensitive data remains within the client’s infrastructure. The new system leveraged machine learning models to detect suspicious transactions, improve accuracy, and identify previously missed fraudulent activity.

Highlights

Machine Learning-Driven Detection

An initial model was trained and validated on known fraud data, achieving an F-Score of 0.84, a strong result by industry standards.

Anomaly Detection Framework

To catch unknown fraud cases, an anomaly detection approach was applied, identifying 8% of transactions requiring further investigation, leading to more comprehensive fraud detection.

Time Series Model (TSM)

A predictive model was developed using historical transaction data to anticipate the characteristics of future transactions. If a transaction deviates significantly from the predicted pattern, it is flagged as suspicious and flagged for review. This also facilitated dynamic transaction limit adjustments.

Results

  • Improved Fraud Detection Accuracy: The new system significantly reduced false positives by analyzing both known and unknown fraud cases.
  • Complete Data Privacy: The on-premises system ensures sensitive transaction data never leaves the client’s infrastructure, addressing full compliance with security and privacy regulations.
  • Dynamic Fraud Monitoring: The Time Series Model introduced the ability to dynamically adjust transaction limits and predict unusual patterns, enabling proactive fraud detection.