Projecting non-conformity scores onto the coherent subspace of a hierarchy makes split-conformal prediction regions smaller while preserving finite-sample coverage, under i.i.d. data and, for the efficiency results, elliptical residual assumptions.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Conformal Prediction for Hierarchical Data
Projecting non-conformity scores onto the coherent subspace of a hierarchy makes split-conformal prediction regions smaller while preserving finite-sample coverage, under i.i.d. data and, for the efficiency results, elliptical residual assumptions.