AUC-based Selective Classification
This addresses a specific need in applications like credit scoring where ranking metrics are crucial, representing an incremental advance in selective classification.
The paper tackles the problem of selective classification for binary classifiers when performance is measured by AUC rather than accuracy, proposing a model-agnostic method called AUCROSS that trades off coverage for AUC, with experiments showing it improves over existing methods.
Selective classification (or classification with a reject option) pairs a classifier with a selection function to determine whether or not a prediction should be accepted. This framework trades off coverage (probability of accepting a prediction) with predictive performance, typically measured by distributive loss functions. In many application scenarios, such as credit scoring, performance is instead measured by ranking metrics, such as the Area Under the ROC Curve (AUC). We propose a model-agnostic approach to associate a selection function to a given probabilistic binary classifier. The approach is specifically targeted at optimizing the AUC. We provide both theoretical justifications and a novel algorithm, called AUCROSS, to achieve such a goal. Experiments show that our method succeeds in trading-off coverage for AUC, improving over existing selective classification methods targeted at optimizing accuracy.