Add Sub-Classifiers Screen Reference



Field
Description
Structure

Select the Structure type the scan should check (e.g. database or system instance).

  • Unstructured and Structured

  • Only Structured

  • Only Unstructured

Minimum Threshold

Define a minimum threshold for classified results from structured data sources to only return classifiers that match the minimum number of occurrences.

The minimum threshold of all classifiers is set at 30% by default. The range can extend from 0 to 100.

Enter Classifier Name

Enter a name to describe the classifier.

Important: Enable the classifier to be active in the group.



Confidence Score

Enter a percentage which will be used to apply the data sample that matches with the discovery pattern. This will also determine the priority level for when the classifier needs to be used. The higher the confidence score, the more precedence the label takes.

Discovery Patterns

Select a discovery pattern and its respective variable/value. Click the Add icon to stack additional discovery patterns. For more information on discovery patterns, see Configuring Discovery Patterns for Classifiers.

  • Mask: checks for a particular alphanumeric sequence (e.g. license number, phone number).

  • Data Type: checks for regularly used data formats.

  • Length Check: checks for a specific character count (e.g. 16 digits for a SSN).

  • Lookup: checks for a specific phrase or term to match against the classifier's metadata (e.g. given names for first name classifier).

  • Regular Expression: enter a regular expression value that aligns with desired pattern.

  • Date: checks for a date range.

  • Digital Checks: checks for a predicate to perform digital checks, valid numbers, and help reduce false positives.

    Note:

    The Custom Digital Check pattern supports a formula builder. For more information, see Configuring Discovery Patterns for Classifiers.

  • NER Model/Athena AI: offers an alternative discovery pattern when found data is not picked up by another pattern.