Make expert rule- and score-based decisions using the Risk IDS Decision Hierarchy, a hierarchical decisioning process. Update data real-time to influence a response, place a record into queue or launch a workflow.
The Decision Engine uses the concept of a Decision Hierarchy, which is an organisation of an overall decision strategy.
A Decision Hierarchy represents different fraud-management disciplines. There may, for example, be a Transaction Fraud Prevention Decision Hierarchy and an Application Fraud Prevention Decision Hierarchy each decisioning different data.
Decision Campaigns are intended to target specific record types; e.g., Applications. For example, in transaction fraud, there may be a Decision Campaign targeting high-risk countries.
Decision Divisions are a prioritised grouping of rules.
Decision Rules are the rules the Decision Engine will test records against.
Records are only eligible for processing against a Decision Rule if it has qualified when tested against the related Decision Campaign.
The Decision Engine will test records against the Decision Rules in order of their Decision Division priority followed by the Decision Rule priority.
The Decision Hierarchy can be configured to test against all of the rules available or stop as soon as a match occurs for any Decision Rule.
Records can be decisioned and a response provided real-time.
Many rules will share common criteria, such as ‘approved transactions.’ The Decision Engine implements a Decision Hierarchy that will allow for filtering to be performed at the Decision Campaign with the qualifying records being stored in a temporary table in memory. The temporary table will then be used for all subsequent Decision Rule processing. This ensures that the majority of rules are evaluated in the application’s memory, thus being fast enough to return responses real-time.
All rules in Risk IDS are compiled into C inside the database, which ensures superb rule-execution times.
Risk IDS has a comprehensive rule capability and when used in conjunction with Risk IDS, Decision Analytics can satisfy the most demanding real-time rule requirements. The following are some of the generic rules that can be deployed in the Decision Engine:
Prior to activating a rule, the test functionality can execute the rule against the historic data. The statistics returned will detail the number of matches a Decision Campaign or Decision Rule has yielded. Additionally, if the risk records are being updated (for example, for a fraud transaction), false\positive ratios will be presented.
Assuming the test statistics are satisfactory, the rule can be enabled to start creating results only, which will allow reports to be produced on the rule’s real match volume.
Rule matches can drive placement of a customer Relationship into Queues for review by an analyst.
Rule matches can trigger Workflows to be launched, which can perform a series of complex actions from performing database procedures to sending notifications. This allows for unlimited processing to take place because of a rule match. A very common implementation in transaction fraud prevention is for a Workflow to integrate with the core account service system blocking a card before the next transaction.
Every record that is created in Risk IDS is processed against the Decision Engine. When a record is processed, an audit of this processing is stored, detailing information on the Decision Campaigns and Rules executed, which includes the order of execution, the matches yielded and the length of time taken to process each step.
These statistics allow for monitoring of the real-time performance of the Decision Engine so an analyst can see which rules would be more appropriate for Near-time processing. Near-time processes the most intensive rules after the response has been dispatched.
Copyright Risk IDS Limited.
UK Registered Company 05026181.
Update
