427 papers
This paper introduces a Rao-Blackwellization-based variance-reduction framework for event-based Monte Carlo probabilistic seismic hazard analysis that eliminates unnecessary ground-motion residual sampling to produce more stable and efficient hazard estimates while unifying the method with classical formulations and extending it to adaptive importance sampling.
This simulation study demonstrates that the Relative Belief Ratio (RBR) filter generally outperforms other feature selection methods in recovering informative binary features and maintaining selection stability when sample sizes are at least 100 and marginal location signals are clear, though its performance declines under heavy-tailed or bimodal distributions where methods like Information Gain or Boruta are more effective.
This paper introduces a selective estimator for sparse circular orientation data that certifies winding residues by enforcing strict smoothness, magnitude, and branch margin criteria, thereby prioritizing the validation of arithmetic consistency over sampling accuracy and frequently abstaining from reporting results when these conditions are not met.
This paper proposes a graph-based audit protocol that distinguishes between level shifts and shape changes to ensure the administrative validity and compatibility of longitudinal datasets joined across different institutional regimes, demonstrating its efficacy through an analysis of Tunisia's economic transition from 2000 to 2019.
This paper conducts a sensitivity audit comparing national-IQ and harmonized-learning indicators, finding high correlations and similar rank distributions but concluding that the datasets are not interchangeable or causally equivalent despite their statistical similarities.
This paper establishes a mathematically rigorous, operator-theoretic framework for the Delphi method that proves epistemic validity depends on specific structural properties of the consensus process, demonstrating through computational experiments that conventional agreement metrics often mask "model collapse" phenomena where superficial consensus coexists with catastrophic loss of legitimate disagreement and increased error.
This study demonstrates that excluding outlier clusters at baseline and adjusting for baseline outcomes can significantly improve the efficiency of malaria cluster-randomised trials, particularly in settings where cluster-level outcomes exhibit strong temporal persistence.
This meta-epidemiologic study of 157 phase III breast cancer trials reveals that survival conclusions often lack statistical robustness, as the number of patients lost to follow-up frequently exceeds the fragility threshold, suggesting that statistical significance alone may overestimate the stability of treatment effects.
This paper argues that distinguishing genuine zoonotic spillover trends from surveillance bias requires robust infrastructure rather than just modeling, demonstrating through simulations and a rabies case study that while trend direction can often be recovered, magnitude estimates are highly unreliable and model disagreement should be treated as a critical signal for reporting.
This study utilizes OpenAlex data to analyze the global scientific citation network, revealing that retracted research is disproportionately concentrated in the core of the largest connected component and exhibits higher structural exposure, particularly in fields like biochemistry and genetics, suggesting that retracted works are more centrally embedded and cross-community connected than non-retracted works.