Install
FixOut is pip-installable and usable directly on your Python IDE or jupyter notebooks. You can also clone it directly from the Github repository.
pip install fixout
For further information on how to install FixOut, please check out the Installation Guide available in the documentation.
Entry points & Components
FixOut has two classes that when combined create an entry point. These classes are Runner (FixOutRunner ☍) and Artifact (FixOutArtifact ☍).
from fixout import FixOutArtifact
from fixout import FixOutRunner
Once the data and model are ready to be analysed, you must encapsulate them in an artifact (object of FixOutArtifact).
fxa = FixOutArtifact(model=model,
training_data=(X_train,y_train),
testing_data=[(X_test,y_test,"Test")],
features_name=features_name,
sensitive_features=["statussex"])
Then use this artifact as an argument to start the runner (object of FixOutRunner).
fxo = FixOutRunner("Fairness Assessment")
fxo.run(fxa)
Once FixOutRunner has finished its execution, other methods can be used to check the calculated values of some fairness metrics, for example.
If you are looking for more details on specific functionalities, check out the API reference or the available examples and tutorials.
FixOut 