When does conformal calibration need censoring weights? Cause-of-failure prediction sets under competing risks

Key points

• Complete-case calibration guarantees coverage for the label-complete subpopulation, but its population coverage can deviate in either direction, even at the true class probabilities.1
• We study how selection changes the score distribution near the population quantile.1
• In these designs, weights from a correctly specified censoring model keep coverage near nominal.1
• Extending the argument of Yi et al. (2025) to a cause label, we establish a finite-sample coverage lower bound with an explicit penalty for censoring-model error.1

Sources (1)

• [1]When does conformal calibration need censoring weights? Cause-of-failure prediction sets under competing risks
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 04:06 PM

  Split conformal prediction sets for competing-risks labels at a fixed horizon require calibration labels that right censoring can leave unobserved.

  Complete-case calibration guarantees coverage for the label-complete subpopulation, but its population coverage can deviate in either direction, even at the true class probabilities.

Extractive summary: sentences quoted from the sources.