Browse Prior Art Database

Transaction Fraud Alert Consolidation Methodology for Financial Institutions Disclosure Number: IPCOM000245829D
Publication Date: 2016-Apr-13
Document File: 6 page(s) / 266K

Publishing Venue

The Prior Art Database


This article discloses a methodology to consolidate alerts that are generated by fraud detection systems in financial institutions.It contains three key ingrdients: (1) an algorithm to decide when alerts needs to be merged (2) a process for merging related alerts (3) an approach to provide an holistic view of the consolidated alerts.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 27% of the total text.

Page 01 of 6

Transaction Fraud Alert Consolidation Methodology for Financial Institutions

Transaction fraud is globally, a crucial problem in financial institutions. Financial institutions throughout the world are actively building counter fraud systems at various levels of sophistication to detect fraud. Such fraud detection systems use statistical analysis and traditional rules to compute a risk rating for transactions performed by an individual. The risk rating is further used to decide if an alert needs to be created for a fraud assessment. For example if the risk rating is high or medium , an alert may be generated. When there is an alert generated, typically a case needs to be created in the case management system for investigation.

This leads to the following problems:

1. Every alert resulting in a case being created is causes too many cases in the case management system thus increasing the operational costs for financial institutions to

perform fraud investigation.

2. Related alerts are being spread across different cases, thus making it difficult to identify

patterns of fraudulent transactions. This is impacting fraud investigation negatively, letting fraudsters escape and eventually leading to increased loss due to fraud.

3. Further, related alerts being investigated by different investigators is leading to an overall increase in the time to investigate potentially fraudulent alerts when compared to a single investigator looking at all closely related alerts. This not only delays fraudsters from being blocked but also causes genuine customers to wait for long before their genuine transactions are allowed to be performed. In other words there is an impact on the customer satisfaction.

Disclosed is a method that attempts to solve the above mentioned problems through identification and consolidation of related alerts and provision of a holistic view of the related alerts to an investigator for further fraud analysis. This solution comprises of :

1.) Alert Consolidation Decisioning (ACD) Algorithm: An algorithm that suggests whether a transaction fraud alert needs to be consolidated with an existing case in the case management system or if a new case needs to be created for this alert for investigation.

2.) Alert Consolidation Process (ACP) that describes logically, how different alerts need to be consolidated into a single case for investigation.

3.) Consolidated Alert Investigation View (CAIV) that provides an approach to present the consolidated alert information to a case investigator for fraud analysis.

When a new alert is generated by the fraud detection system, the alert attributes are passed on to the proposed algorithm as input. The algorithm looks into existing cases in the case management repository and outputs whether a new case needs to be created for the new alert in context or if the alert needs to be consolidated with an existing case in the case management system.


Page 02 of 6

If the algorithm decides that a consolida...