Reducing False Positives to Improve Regulatory Reporting for a Global Bank
Overview.
A global bank faced challenges in adhering to regulatory compliance and reducing false positives in anti-money laundering (AML) activities. Coforge implemented a machine learning solution that utilized soft clustering and behavioral misalignment scoring, leading to significant improvements in fraud detection rates, efficient fraud management, and effective workload management.
Anomalous Behavior Monitoring: Needed to monitor compliance databases like OFAC to catch anomalous behaviors.
High False Positives: Existing transaction rules generated many false positives due to evolving behaviors.
Our Solution.
Coforge implemented a comprehensive solution to address the client's challenges:
Soft Clustering Algorithm: Used a behavioral misalignment score to detect anomalies.
Customer Clustering: Aggregated banking transactions and built customer clusters to identify suspicious behaviors.
Suspicious Customer Identification: Identified suspicious customers who were not in any cluster, misaligned within clusters, or deviated from cluster prototypes.
ML Classification Prediction Model: Used historical false positive alerts and related information as input to predict new alerts.
Compliance Database Checks: Employed fuzzy and regex matching techniques to check against compliance databases.
Classification Score: Applied the model to new inputs to generate a classification score.
Key Highlights:
Coforge's innovative solution delivered significant value to the client's operations:
Reduced False Positives: Achieved a significant reduction in false positives.
