Applying Artificial Intelligence to Detect Hidden Data Anomalies

Data integrity issues within capital markets can long remain undetected before surfacing through inconsistencies in the dataset.

To support the quality and reliability of information uploaded to its database, European DataWarehouse (EDW) performs a range of verification checks designed to identify potential errors, anomalies and inconsistencies.

This whitepaper explores how Isolation Forest, an unsupervised machine learning technique, could complement these existing controls by identifying unusual observations that may warrant further investigation.

In this whitepaper, readers will discover:

• How Isolation Forest can be applied to loan-level securitisation data
• The potential of machine learning to support anomaly detection and quality assurance processes
• Insights from a practical application using UK RMBS loan-level data
• Opportunities to enhance data quality monitoring without additional reporting requirements
  

EDW is uniquely positioned to support such efforts given its extensive repository of standardised loan-level data across a large universe of public European securitisation transactions.

Learn how EDW applies machine learning to identify hidden data anomalies and strengthen data quality monitoring. Download the whitepaper below.