Halfaker, an SAIC company, has a government client that created a fraud detection system to aggregate and scan transaction-level data to ensure retailer compliance and identify suspicious retailer activity of critical food assistance programs. This system synthesizes data in a combination of reports, queries, and visualizations to enable the client to determine suspect transactions that represent program violations, playing a critical role in effectively monitoring, investigating, and prosecuting against fraudulent retailers. With ever-evolving trends in fraudulent activities, our client needs continuous enhancements and fraud detection capability development in order to ensure the integrity of its programs, maximize program benefits to recipients, and minimize operational risks.

The challenge

Given evolving trends in fraudulent activities, our government client’s fraud detection system requires continuous enhancements and development in fraud detection capabilities to ensure the integrity of the its assistance program, maximize the benefits to aid recipients, and minimize operational risks.

Our solution

Halfaker leverages data mining best practices to develop supervised machine-learning models by applying the cross-industry standard for data mining (CRISP-DM) methodology and industry-leading technologies, including R and Tableau. Our data scientists leverage historical case data in combination with statistical methods, mathematical formulas, models, and techniques to analyze data and illuminate potentially fraudulent behavior based on known trends and suspicious data patterns. Our approach to big data analytics provides the client with industry-leading predictive modeling, machine learning, data mining, large database administration, data handling, data validation, and database application development support its fraud detection system.

Our iterative agile development approach aligns with the client’s SDLC requirements and fosters collaboration, promotes adaptive planning, and encourages rapid and flexible enhancements to the fraud detection system. As features and services are added, we leverage a proactive ITIL-based approach to provide ongoing operations support and conduct ongoing monitoring to improve application, server, and performance; enabling us to exceed the 99.8% system availability performance requirement. To increase the our client’s security posture, we use SolarWinds for continuous monitoring, intrusion detection, and critical infrastructure protection to ensure security compliance with client and federal standards (FISCAM, NIST, FISMA), detect system failures at an individual process level, and provide threat intelligence on potential malicious activity.

Realized benefits

Halfaker’s software engineering solution improves user experience and minimizes operational risk to the client’s fraud detection system, while our innovative fraud detection solution helps to ensure food assistance program integrity. To date, the client has achieved a 60% increase in fraud detection and disqualified over 3,500 retailers from its food assistance program, increasing program efficacy and saving American tax dollars. The client also has benefited from:

• Enhanced usability with mobile-friendly design
• Improved data quality
• Proactive fraud detection