COMPAS Dataset

The COMPAS recidivism dataset contains 6,167 samples, and the task is to predict the recidivism of an individual based on criminal history, with the binary protected attribute being race (white and non-white) and an additional nine features.