Reliable classification through rank-based conformal prediction sets

Rui Luo, Zhixin Zhou

Abstract: Highlights•A novel conformal prediction method based on a rank-based conformity score function, suitable for classification models that may not output well-calibrated probabilities.•Theoretical analysis of the coverage guarantee and the expected size of the conformal prediction sets based on the rank distribution of the underlying classifier.•Extensive empirical evaluation demonstrating the effectiveness of our method in providing reliable uncertainty quantification for classification tasks across various domains.

External IDs:doi:10.1016/j.patcog.2025.112330