I'm looking to build a classifier that can label documents as private or public based on their content and need to find a suitable dataset that contains private/confidential/secret/sensitive documents.

It's fairly easy to understand that they are hard to find since they contain 'private/confidential/secret/sensitive' information. However, most datasets I have discovered have had the important parts redacted or cleansed.

Anyone have any suggestions for a dataset or even a place to look for such a dataset?

  • something like this? neboagency.com/blog/…
    – philshem
    Commented Sep 20, 2019 at 18:06
  • private or public based on their content -> What's private depends on context (E.g. 1, 2), so if you are not defining any context it's practically impossible what private/etc is (other than those exact words at the top of the text). Please edit your question to give context.
    – user4293
    Commented Sep 25, 2019 at 14:14


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.