Identifying Significant Predictive Bias in Classifiers

Zhe Zhang, Daniel B. Neill

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, a…