427 papers
This paper demonstrates that in spatially heterogeneous systems like urban migration, feature importance rankings derived from machine learning are unstable across different development stages and do not inherently reveal the direction or uniformity of a predictor's influence, necessitating separate spatial validation for policy applications.
This paper introduces the Finite-Evidence Decision Identification Procedure (FEDIP), a modular audit framework that reveals how finite validation data often fails to uniquely identify a single optimal action among competing distributional models, thereby demonstrating the necessity of explicit economic loss functions to resolve action dispersion that standard score-based selection overlooks.
This study introduces and validates a Gaussian-based refinement algorithm that successfully estimates population usual energy intake distributions from Food Balance Sheets by treating reported values as upper bounds, demonstrating significantly improved accuracy over existing FAO methods when compared to national dietary surveys across six countries.
This study empirically compares linear mixed-effects, generalized estimating equations, and two-part mixed models for estimating EQ-5D utilities in rheumatoid arthritis trials, finding that while two-part mixed models yield lower utility estimates, the choice of method has minimal impact on overall cost-effectiveness conclusions.
This paper proposes a unified methodological framework combining a Global Normative-Anchor Translated CES function with a certainty-equivalent pre-screening mechanism to resolve computational domain collapse and prevent the "Epistemic Shielding Trap" in non-compensatory composite indicators, thereby enabling accurate longitudinal data quality audits while explicitly warning against using the penalized index for cross-sectional funding allocation.
This paper introduces a Bayesian modeling framework that jointly addresses misclassification and selection errors in multi-source binary data, demonstrating through simulations and empirical analysis that while composite error can be robustly identified, accurately decomposing it into specific components requires valid indicators, auxiliary variables, and informative priors.
This paper proposes a general framework for quantifying the stability of non-deterministic prediction models across metric, categorical, and ordinal outcomes by introducing data-based and model-based stability measures that overcome the limitations of existing methods, such as their inability to handle individual test objects and their sensitivity to data composition.
This study develops a sequential tour-based mode choice framework for Sydney using GPS data and machine-learning-derived features, demonstrating that Mixed Logit models significantly improve prediction for interdependent tour components and that the framework remains robust to upstream estimation uncertainty, thereby validating the use of simulated inputs for practical travel demand forecasting.
This study demonstrates that on credit default prediction tasks, XGBoost's built-in `scale_pos_weight` parameter significantly outperforms the traditional SMOTE resampling technique in recall, highlighting that the optimal strategy for handling class imbalance depends on the specific classifier used rather than a universal resampling method.
This paper proposes a method for constructing estimated ranking tables with joint confidence regions for the true ranking of K populations, utilizing normality assumptions and standard errors to visually communicate ranking uncertainty and provide marginal confidence sets for individual ranks.