[FEATURE] Fairness algorithm that does not require the test dataset to contain sensitive features

[kimeetok]

Is there an existing issue for this?

• I have searched the existing issues

Description

I would like a testing algorithm for fairness that does not require the test dataset to contain sensitive features as my model was not trained with any sensitive features.

Is your feature request related to a problem/limitation?

This is in relation to #226

Describe the solution you'd like

This alternative fairness test could probably take in an extra dataset as a test argument.
This extra dataset will contain any indirect sensitive features that correspond with each row in the testing dataset.

Alternatives

No response

Additional Context

No response

Related Issues/PRs

#226