The document parser extracts entities bases on the entity name and aliases.
However a lot of aliases are very generic, and not suitable for extraction.
For example, some malware has an alias "Agent",
Industry sectors have aliases such as "research", "software", ...
Almost all reports have words like this, and it creates a lot of entities that are not related to the report.
It would be nice if we could keep aliases, but disable some for extraction
Current Workaround
Deleting the aliases, but this removes information..
Proposed Solution
Provide a property on the alias, to indicate that it shouldn't be used for entity extraction
If the feature request is approved, would you be willing to submit a PR?
Use case
The document parser extracts entities bases on the entity name and aliases. However a lot of aliases are very generic, and not suitable for extraction. For example, some malware has an alias "Agent", Industry sectors have aliases such as "research", "software", ... Almost all reports have words like this, and it creates a lot of entities that are not related to the report. It would be nice if we could keep aliases, but disable some for extraction
Current Workaround
Deleting the aliases, but this removes information..
Proposed Solution
Provide a property on the alias, to indicate that it shouldn't be used for entity extraction
If the feature request is approved, would you be willing to submit a PR?
Not at the moment