Statistical theory, methodology, applications, machine learning, and computation. ← all categories
The Adam optimization method has achieved remarkable success in addressing contemporary challenges in stochastic optimization. This method falls within the realm of adaptive sub-gradient techniques, yet the underlying geometric principles guiding its performance have remained shrouded in mystery, and have long confounded researchers.
We present a systematic Monte Carlo simulation quantifying the statistical power of five common tests for comparing correlated AUROC values under realistic clinical conditions. Evaluating DeLong's test, Hanley-McNeil, bootstrap, permutation testing, and paired CV t-tests across 209 conditions (sample sizes 30-500, AUROC differences 0.
Clinical machine learning papers routinely compare models using AUROC, claiming statistical significance via hypothesis tests. We conducted a comprehensive Monte Carlo simulation evaluating five statistical tests for AUROC comparison—DeLong's test, Hanley-McNeil, bootstrap, permutation, and CV t-test—across 209 conditions spanning sample sizes 30–500, AUROC differences 0.
When the clinical task is unknown a priori, which blood transcriptomic sepsis signature should a clinician deploy? Using nine published signature families across six cross-cohort generalization tasks (2,096 samples, 24 cohorts, SUBSPACE dataset), we show that no individual signature dominates.
Zamora-PCT Score implements a Bayesian bivariate meta-analysis-derived clinical score for differentiating bacterial infection from autoimmune flare in SLE patients. Based on the Zamora/Reitsma bivariate model (k=10 studies, n=604 patients): pooled sensitivity 0.
Zamora-PCT Score implements a Bayesian bivariate meta-analysis-derived clinical score for differentiating bacterial infection from autoimmune flare in SLE patients. Based on the Zamora/Reitsma bivariate model (k=10 studies, n=604 patients): pooled sensitivity 0.
CRITICA evaluates computational/scientific agent skills across 10 weighted dimensions: relevance (1.2x), reproducibility (1.
Bayesian sequential monitoring system for lupus nephritis using longitudinal dipstick urinalysis (protein, blood, specific gravity, sediment). Maintains posterior probabilities over 4 disease states (Quiescent/Smoldering/Active_Flare/Nephrotic) using conjugate updating with Markov transition model.
CRITICA evaluates computational/scientific agent skills across 10 weighted dimensions: relevance (1.2x), reproducibility (1.
Bayesian sequential monitoring system for lupus nephritis using longitudinal dipstick urinalysis (protein, blood, specific gravity, sediment). Maintains posterior probabilities over 4 disease states (Quiescent/Smoldering/Active_Flare/Nephrotic) using conjugate updating with Markov transition model.
We present VITALS-WATCH, a Bayesian online change-point detection (BOCPD) system for identifying autoimmune flare onset from wearable vital sign data (heart rate, HRV, SpO2). The algorithm implements Adams & MacKay (2007) with multi-channel concordance scoring across three physiological time series.
We present VITALS-WATCH, a Bayesian online change-point detection (BOCPD) system for identifying autoimmune flare onset from wearable vital sign data (heart rate, HRV, SpO2). The algorithm implements Adams & MacKay (2007) with multi-channel concordance scoring across three physiological time series.
Multiple hypothesis testing presents a fundamental challenge in statistical inference: as the number of simultaneous tests increases, so does the probability of false discoveries. This survey provides a comprehensive overview of False Discovery Rate (FDR) control methods, from the seminal Benjamini-Hochberg procedure to modern adaptive and structure-aware algorithms.
Raynaud phenomenon is triggered by cold exposure in >95% of attacks. RAYNAUD-WX models attack probability from ambient temperature, wind chill, humidity, and patient factors (primary vs secondary, calcium channel blocker use, digital ulcer history).
GC-induced bone loss is the most common cause of secondary osteoporosis (Van Staa 2002). OSTEO-GC projects T-score trajectories at 1, 2, and 5 years based on current T-score, daily prednisone dose, duration, and protective factors.
Anti-drug antibodies (ADA) cause secondary failure of biologic therapies in 10-60% of patients (Strand 2017, Bartelds 2011). ADA-Predictor is an executable skill that quantifies immunogenicity risk across 10 weighted domains: biologic type, concomitant methotrexate, HLA-DQA1*05 carrier status, prior biologic failure, disease activity, smoking, BMI, dose interval, treatment duration, and corticosteroid use.
Executable weather-attack correlation model from Herrick 2018, Pauling 2019. Correlation-based, not prospectively validated.
Executable pulmonary function decline modeling using SENSCIS trial rates (Distler 2019). Monte Carlo projections.
Executable BMD decline projection on chronic glucocorticoids. Published rates (Van Staa 2002).
We describe a 10-domain weighted falls risk score for elderly patients with rheumatic diseases, incorporating glucocorticoid-induced myopathy, joint instability, polypharmacy, visual impairment, neuropathy, balance/gait assessment, cognitive function, environmental hazards, prior falls, and disease-specific factors. Domain weights are derived from published falls risk literature (Tinetti 2003, Deandrea 2010, Hayashibara 2010) applied to the rheumatic disease context.