Empirical Likelihood Preserves Shape in Covariate-Adjusted Trials
A covariate-balanced empirical measure lets randomized-experiment analyses improve efficiency while retaining valid distribution and survival-curve shape.
Aug 24, 20263 min2608.19423
Statistics Theory
math.STApplied and theoretical statistics including point estimation and hypothesis testing.
Sort:
Spectral Certificates Separate Accuracy From Inference
An explicit Matrix-Bernstein and gap protocol gives finite-sample network certificates, then declines when valid sets are still vacuous.
Jul 30, 20265 min2602.10566
Rank-Based Shrinkage Targets Heavy-Tailed Financial Covariance
MENS combines normal-scores ranks with nonlinear eigenvalue shrinkage to estimate covariance in nonparanormal financial models.
Jul 28, 20264 min2607.19825
Mechanistic Tumour Dynamics Extend Oncology Endpoint Forecasting
PIONEER couples latent SLD trajectories to multistate hazards, forecasting PFS at month 4 and OS by month 11 in an ES-SCLC case study.
Jul 28, 20264 min2607.17908
BLASSO Extends Mixture Recovery to Unknown Diagonal Covariances
A continuous sparse estimator recovers Gaussian mixture weights, means, and diagonal scales under explicit separation, with near-parametric prediction rates.
Jul 28, 20265 min2509.12889
Rank-Based Nonlinear Shrinkage Targets Heavy-Tailed Covariance Estimation
MENS estimates high-dimensional latent covariance in nonparanormal models by applying oracle nonlinear shrinkage to normal-scores rank covariance eigenvalues.
Jul 28, 20263 min2607.20576
Residual Tail Mass Breaks Survival Score Properness
A marginal-properness analysis derives the SBS bias term and tests it across 10,000 simulations under low and high censoring.
Jul 28, 20265 min2212.05260
Poincaré Bases Extend Gradient Sensitivity Analysis Beyond Polynomials
A Sturm--Liouville chaos basis makes derivative information orthogonal, improving Sobol index estimation on toy and flood models with small designs.
Jul 13, 20265 min2510.03056
Divergence Estimators Get a Model-Invariant Robustness Rule
A Pareto-frontier analysis of GABD estimators identifies an extended $(\phi,\gamma)$ divergence that minimizes asymptotic variance at a chosen breakdown point.
Jul 13, 20264 min2607.04343