Identifying Significant Predictive Bias in Classifiers
This work addresses the need for interpretable bias detection in classifiers, which is crucial for fairness and accountability in domains like criminal justice and finance, though it is incremental in extending subset scan methods to this context.
The authors tackled the problem of detecting statistically significant predictive bias in probabilistic binary classifiers, presenting a novel subset scan method that identifies biased subgroups across all possible feature combinations, and demonstrated its application on COMPAS recidivism and credit delinquency datasets.
We present a novel subset scan method to detect if a probabilistic binary classifier has statistically significant bias -- over or under predicting the risk -- for some subgroup, and identify the characteristics of this subgroup. This form of model checking and goodness-of-fit test provides a way to interpretably detect the presence of classifier bias or regions of poor classifier fit. This allows consideration of not just subgroups of a priori interest or small dimensions, but the space of all possible subgroups of features. To address the difficulty of considering these exponentially many possible subgroups, we use subset scan and parametric bootstrap-based methods. Extending this method, we can penalize the complexity of the detected subgroup and also identify subgroups with high classification errors. We demonstrate these methods and find interesting results on the COMPAS crime recidivism and credit delinquency data.