# Logistic Regression

Clinical prediction models inherit their quality from routine-care training data, yet missingness rates and class ratios are not tied to what happens after deployment. An exploratory benchmark across three MIMIC-IV cohorts and the multi-institution eICU database splits data quality into completeness, label noise, distributional similarity, representativeness and temporal integrity, training a real model under each controlled condition and measuring held-out log-loss and calibration change. The composite moved in the expected direction in all three controlled cohorts, but combined corruptions were sub-additive rather than multiplicative, and across 21 analyzable eICU hospitals it tracked calibration but not detectably discrimination.

Clinical prediction models inherit their quality from routine-care training data, yet missingness rates and class ratios are not tied to what happens after deployment. An exploratory benchmark across three MIMIC-IV cohorts and the multi-institution eICU database splits data quality into completeness, label noise, distributional similarity, representativeness and temporal integrity, training a real model under each controlled condition and measuring held-out log-loss and calibration change. The composite moved in the expected direction in all three controlled cohorts, but combined corruptions were sub-additive rather than multiplicative, and across 21 analyzable eICU hospitals it tracked calibration but not detectably discrimination.

Many variants found in monogenic diabetes genes remain variants of uncertain significance because evidence is insufficient or conflicting. A pipeline trained on 7,706 confidently classified single nucleotide variants across 22 such genes reached ROC-AUC 0.9989 and AUPRC 0.9970 on 771 held-out variants, with recall 1.00 on the pathogenic class. Feature ablation showed that part of this accuracy reflects overlap with the ACMG criteria used to assign ClinVar labels, yet on non-missense variants, where computational evidence could not have been directly invoked, the model still beat a majority-class baseline, 99.0% versus 87.0%; it then ranked 6,306 uncertain variants with per-variant explanations.

Many variants found in monogenic diabetes genes remain variants of uncertain significance because evidence is insufficient or conflicting. A pipeline trained on 7,706 confidently classified single nucleotide variants across 22 such genes reached ROC-AUC 0.9989 and AUPRC 0.9970 on 771 held-out variants, with recall 1.00 on the pathogenic class. Feature ablation showed that part of this accuracy reflects overlap with the ACMG criteria used to assign ClinVar labels, yet on non-missense variants, where computational evidence could not have been directly invoked, the model still beat a majority-class baseline, 99.0% versus 87.0%; it then ranked 6,306 uncertain variants with per-variant explanations.

Misclassifications from power-grid intrusion detection are converted into generator-level cyber-physical risk costs that enter dispatch optimization, instead of optimizing detection F1 alone. Previously detection and dispatch were optimized separately, so the operational cost of detection errors never entered the decision objective. On the ICS and Sherlock grid datasets, after a grid search over false-positive, false-negative and risk-scaling parameters, some configurations produced lower minimum dispatch costs than their baselines, with the size of the improvement varying by dataset and configuration.

Misclassifications from power-grid intrusion detection are converted into generator-level cyber-physical risk costs that enter dispatch optimization, instead of optimizing detection F1 alone. Previously detection and dispatch were optimized separately, so the operational cost of detection errors never entered the decision objective. On the ICS and Sherlock grid datasets, after a grid search over false-positive, false-negative and risk-scaling parameters, some configurations produced lower minimum dispatch costs than their baselines, with the size of the improvement varying by dataset and configuration.

This work characterizes the geometry of hidden-state probe signals for LLM hallucination detection across three 7B-scale models and three datasets in a paired-example paradigm, finding the signal overwhelmingly dominated by a single mean-shift component whose removal collapses detection to chance, with shrinkage linear discriminant analysis closing about 73% of the gap between 1D and full-dimensional classifiers, a simple L2-regularized logistic regression at 0.952 AUROC bounding or outperforming twelve controlled architectural alternatives, and LayerMix multi-layer aggregation matching oracle-layer performance without oracle access.

This work characterizes the geometry of hidden-state probe signals for LLM hallucination detection across three 7B-scale models and three datasets in a paired-example paradigm, finding the signal overwhelmingly dominated by a single mean-shift component whose removal collapses detection to chance, with shrinkage linear discriminant analysis closing about 73% of the gap between 1D and full-dimensional classifiers, a simple L2-regularized logistic regression at 0.952 AUROC bounding or outperforming twelve controlled architectural alternatives, and LayerMix multi-layer aggregation matching oracle-layer performance without oracle access.

Using two prospective cohorts, the China Health and Retirement Longitudinal Study (CHARLS, 2011–2018, n=3,622) and the English Longitudinal Study of Ageing (ELSA, 2012–2020, n=2,059), this study followed middle-aged and older adults with LAP-defined MASLD and no baseline frailty, selected 14 consensus predictors from 31 candidates via LASSO, Boruta, and recursive feature elimination, and trained nine machine learning models; logistic regression performed best (internal test AUC 0.740; external validation AUC 0.753), pain was the top predictor (mean absolute SHAP 0.184), pain remained associated with incident frailty after full adjustment including baseline frailty index (CHARLS RR 1.219; ELSA RR 1.310) with PAFs of 8.01% and 11.

Using two prospective cohorts, the China Health and Retirement Longitudinal Study (CHARLS, 2011–2018, n=3,622) and the English Longitudinal Study of Ageing (ELSA, 2012–2020, n=2,059), this study followed middle-aged and older adults with LAP-defined MASLD and no baseline frailty, selected 14 consensus predictors from 31 candidates via LASSO, Boruta, and recursive feature elimination, and trained nine machine learning models; logistic regression performed best (internal test AUC 0.740; external validation AUC 0.753), pain was the top predictor (mean absolute SHAP 0.184), pain remained associated with incident frailty after full adjustment including baseline frailty index (CHARLS RR 1.219; ELSA RR 1.310) with PAFs of 8.01% and 11.