Researchers have introduced a novel robust conformal prediction method designed to handle datasets with outliers or heavy tails. This new approach utilizes a non-conformity score based on the half-mass radius, which is equivalent to the distance to the $(\lfloor n/2\rfloor+1)$-nearest neighbor. The method ensures marginal validity for conformal regions across all sample sizes and converges to a robust population central set, providing theoretical guarantees for its performance with challenging data distributions. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
RANK_REASON The submission is an arXiv preprint detailing a new statistical method for conformal prediction.
