427 papers
This paper introduces DriftGuard, a two-stage auditing framework that establishes theoretical detectability boundaries for distribution shift and provides a rigorous, label-budgeted protocol to diagnose coverage failure and recalibrate conformal prediction models under both covariate and concept shifts.
This study demonstrates that exposure-linked differential recording of continuous outcomes, such as visually estimated blood loss, can fabricate large, robust-looking treatment effects in real-world evidence through systematic measurement error rather than true clinical differences, necessitating specific diagnostic checks and reliance on objective measures to avoid spurious conclusions.
This paper presents an uncertainty-aware, spatially informed mixture model with shrinkage and class-weighting mechanisms that significantly improves the classification of imbalanced CO₂ plume states in the Sleipner benchmark by identifying interpretable depth-organized regimes and providing posterior uncertainty estimates.
This paper proposes Bayesian kernel machine meta-regression (BKMMR), a flexible second-stage modeling framework that overcomes the limitations of conventional linear meta-regression by automatically capturing nonlinearities and interactions in multi-location environmental epidemiology data, thereby improving estimation accuracy, prediction, and downscaling capabilities as demonstrated through simulations and an air pollution study in South Korea.
This study demonstrates that the Right-Truncated Peaks-Over-Threshold (RT-POT) model outperforms traditional Generalized Extreme Value and Generalized Pareto Distribution approaches by providing more precise parameter estimates and realistic upper endpoints for managing extreme water levels at Ghana's Akosombo Dam under natural truncation.
This paper proposes Adaptive Bayesian Threshold Heuristic strategies for the partial-information secretary problem by integrating full-information optimal stopping theory with Bayesian updating via a Normal-Gamma conjugate prior, demonstrating superior performance over maximum likelihood estimation methods, particularly under small sample sizes and weak prior information.
This paper proposes and evaluates the Cross-Fitted Contamination-Aware Generalized Empirical-Bayes Liu Shrinkage (CF-CABLS-MNL) estimator for Multinomial Logit models, a robust framework that integrates cross-fitted density-power-divergence, misclassification adjustment, and spectral shrinkage to significantly reduce mean squared error under multicollinearity and outliers, though empirical results indicate that while it excels in contaminated settings, direct regularization methods like Ridge regression may outperform it in clean or dense-data scenarios.
This paper proposes a model-based clustering framework for spherical and hyper-spherical data using elliptically symmetric distributions, specifically the elliptically symmetric angular Gaussian and spherical elliptically symmetric projected Cauchy distributions, which are estimated via the expectation–maximization algorithm and evaluated through simulations and real-world applications.
This study of 2,058 matches across six European leagues demonstrates that while relative performance indicators (team–opponent differences) are more effective than absolute indicators in predicting match outcomes, the most accurate predictive model is achieved by integrating both absolute and relative metrics.
This paper introduces the two-parameter Odd Log-Logistic Rayleigh (OLLR) distribution as a flexible extension of the Rayleigh model, deriving its statistical properties and demonstrating its superior performance in modeling lifetime data through applications on two real-world datasets.