Linked Analysis lets you compare relationships between different datasets (for example, billing email, IP address, billing phone) to identify patterns and inconsistencies. With this analysis, you can locate data to updates lists, identify trends, and manually review orders for fraud. When performing a linked analysis, you can select multiple datasets to compare across orders in Event Analysis. If the datasets are linked to other orders, those orders are displayed in a table and the linked information is highlighted.
While the functionality is the same, you can use Linked Analysis in Case Management and Event Analysis for different use cases.
Linked Analysis in Case Management
Linked Analysis in Case Management is primarily used for reactive investigation on a single transaction that has been flagged for manual review. The analyst’s goal is to make a final, educated decision on the specific order.
Use cases
-
Manual reviewing orders for fraud: This is the core function, where an analyst traces the connections of a high-risk order to quickly determine if it is tied to previously confirmed fraudulent activity.
-
Locating data to update lists: Based on the confirmed links to other transactions, the analyst can identify specific data points (like billing email or IP address) to update lists (like VIP or Decline lists) to ensure future orders are handled correctly.
Linked Analysis in Event Analysis
Linked Analysis in Event Analysis is used for a proactive, strategic approach. The focus is on large-scale data exploration to understand emerging fraud patterns and tune system-wide policies.
Use cases
-
Identify trends: Analysts select multiple datasets to compare across orders in Event Analysis to spot new or emerging fraud rings and attack patterns across the entire pool of recent transactions.
-
Pattern and inconsistency identification: Used to identify system-wide patterns and inconsistencies by seeing how widely data points (like a common IP address or phone number) are being shared across many transactions.
-
Policy and rule refinement: The trends and patterns identified here are used to refine and optimize the automated policies and rules that govern fraud decisions for all incoming orders.
Choosing a matching logic (AND/OR) allows you to control how your data connects. Whether you need to broaden your search to find related activity or tighten your focus to uncover specific, high-confidence clusters, this feature provides the precise control required for in-depth fraud investigations. You can select Or logic for broad initial discovery across multiple data points, or select And logic to isolate precise clusters that match every selected data point.
To perform a linked analysis:
-
Sign in to Kount 360.
-
Select Event Analysis or Case Management.
-
Find an order, then select the order to open order Details.
-
In the Linked Analysis widget, select the datasets you want to compare.
-
Select a logic for the linked fields. Choose Or logic for broad initial discovery across multiple data points, or choose And logic to isolate precise clusters that match every selected data point.
-
Select Run Link Analysis.
The Link Analysis table displays with the linked orders highlighted by row and column.
Note
In Linked By, you can select or deselect datasets to visualize more linked datasets in the table.
Comments
0 comments
Article is closed for comments.