Pseudonym Management
Configure the retention period for pseudonyms and sessions.
In pseudonymization mode, the technology automatically generates a pseudonym for each new entity detected. The syntactic variants of the same entity are assigned the same pseudonym (see grouping). Pseudonyms are persistent throughout the session.
Pseudonymization Table
Each session maintains a table that associates each original value with its pseudonym, accessible from the Pseudonyms tab of each session. The table is automatically populated as new entities are detected and retains the mappings for:
- Consistency within the document and across documents for better analysis
- Subsequent restoration, which can be performed directly from the session’s Restoration tab
Pseudonym Grouping
The technology gathers every variation of the same entity under one pseudonym, including the differing syntaxes and the spelling errors (Aurélie Lefevre, LEFEVRE Aurélie, Aurélie Lefebvre).
Pseudonym Variants
We discussed earlier the concept that a single entity can have different syntactic variants: Aurélie Lefevre, LEFEVRE Aurélie, Madame Lefevre, etc. (including variants containing spelling errors).
These different variants can be handled in one of the following ways:
- Merged: All variations are therefore grouped under a single label, without distinction. This makes the text easier to read, but when restoring the text, the reference value is systematically restored in place of the pseudonym, regardless of the variant.
- Separated: All variations are grouped under a single label, to which a suffix is added to distinguish each variation (e.g.,
PR1aforAurélie Lefevre,PR1bforMadame LEFEVRE,PR1cforAurélie, etc.). The letters range from a to z, then aa, ab…, and are applied consistently throughout the entire session. De-pseudonymization therefore restores the exact variant of each sub-pseudonym.
The Pseudonym Variants setting is configured for each session from its Overview tab, or by default from the Processing page:
Editing a Pseudonym
From the Pseudonyms tab in each session, by clicking the settings icon for an alias, you can:
- correct the boundaries of an incorrectly detected entity, for example, when a text fragment has been aggregated into it
- merge or split pseudonyms
- set a new reference value
Pseudonym Retention
Did you know you can configure how long pseudonyms are retained? Go directly to the Privacy page to select your preferences.
By default, documents within a session are retained only for the duration of processing (strict non-retention policy).
Sessions and their associated pseudonyms are automatically deleted based on your selected settings. Of course, you can also manually delete sessions directly from the Sessions tab.