Derives analytical characterizations of the length-optimal training-calibration split ratio for split conformal prediction in general settings and specific regression models.
On the rate of convergence of fully connected deep neural network regression estimates.The Annals of Statistics, 49(4):pp
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.ST 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
On Optimal Data Splitting for Split Conformal Prediction
Derives analytical characterizations of the length-optimal training-calibration split ratio for split conformal prediction in general settings and specific regression models.