Example notebook:

FairML library

An easy to use Python Library to train and develop new Machine Learning models within some fairness constraints. This is an implementation of this Science paper.
Also includes some other handy tools like:

• Bound propogation using the RandomVariable object.
• [Documentation WIP] Python implementation of the CMA-ES black-box optimization algorithm. You can refer to the details here and the implementation here

Installation

Currently, you can install the library only from source using pip:

pip install https://github.com/hannanabdul55/seldonian-fairness/archive/master.zip

Usage

Use this notebook as a reference to train a basic Logistic Regression Model.
A quickstart tutorial on how to get quickly get started with developing your own model is present here. Alternatively, you could use the LogisticRegressionSeldonianModel to train a Logistic Regression model with any scipy.optimize.minimize method by specifying it when calling the fit method.