Anomaly Detection

WFP is committed to putting measures in place to analyze the vast amounts of data generated during regular operations of delivering cash assistance to the people we serve. These measures are necessary in detecting fraudulent activities or any other anomalies which are critical in improving system reliability and enhancing operational efficiency.

Anomaly Detection provides detailed insights into workflows to better understand anomalies in WFP CBT operations, global fleet operations and other operations within WFP. The tool automates data collection, pre-processing, model training, and analysis, ensuring WFP maintains data accuracy, improves efficiency and reduces the risk of missed irregularities.

The tool is designed to handle growing amounts of data and improve how we detect anomalies, greatly increasing the efficiency and reliability of key WFP operations.

Anomaly Detection is in its development stage. The team so far has finalized its strategic approach and identified a clear path to the necessary data. In addition, the team has developed a proof of concept anomaly detection model to test the feasibility of the intended solution.