external-validation

AI-powered chest X-ray pneumonia detection with 86% accuracy and 96.4% sensitivity, validated on an independent (cross-operator) cohort of 485 pediatric samples. Built with TensorFlow & FastAPI.

预测模型自学教程：从数据清洗到模型外部验证（R 语言）

Predictions-only calibration validation for binary clinical prediction models: cluster-aware, cross-fitted, report-emitting. Python.

Reproducible TCGA-LUAD survival pipeline on RNA-seq, clinical and mutation data. LASSO-Cox gene signature overfit (train C=0.722, test 0.538), reported as a negative finding; clinical Cox model (age, sex, grouped stage) validated on external cohort GSE31210 (C=0.708). Snakemake, Docker, FastAPI.

Developed ML algorisms on transcriptomics landscapes identi-fies SPON1 and ALDH1A2 as candidate prognostic biomarkers in TME and serous ovarian cancer progression

Reproducible regulatory transfer analysis for CRISPR interference in primary human CD4 T cells

Interpretable TCGA-BRCA RNA-seq breast cancer subtype classification with ML, SHAP, pathway enrichment, and METABRIC external validation.

SpectraLite and NeuroChronoGraph: transferring EEG dementia classifiers and their referral thresholds across six clinical cohorts (code for the IEEE J-BHI submission)

Biomedical knowledge-graph reliability screening: source code, frozen protocol, external-data evaluation and derived evidence. Research software; no journal publication claimed.

Переносятся ли модели выживаемости при раке молочной железы на новую когорту? Обучение на TCGA-BRCA, внешняя проверка на METABRIC: регрессия Кокса, случайный лес выживаемости, C-индекс Уно, калибровка.

Designed and validated a computational oncology pipeline for hepatocellular carcinoma using RNA expression data, integrating machine learning, survival analysis, SHAP explainability, tumor microenvironment deconvolution, and multi-cohort external validation for precision medicine applications.

Analysis code for a laboratory-based frailty index (FI-Lab) in older ICU patients: derivation in MIMIC-IV and frozen-coefficient external validation in NWICU and eICU-CRD

Physiology-led study of a 1D ResNet for 12-lead ECG diagnosis (PTB-XL): accuracy, noise, patient groups, where the AI looks (Grad-CAM), and a second-dataset test (Georgia). Limits reported openly.

Reproducible analysis notebooks for self-report–wearable measurement discrepancies, perceived stress, and external validation using Samsung Health and Fitbit datasets.

Independent external validation, recalibration & fairness audit of EchoNext-Mini (AI-ECG for structural heart disease) in MIMIC-IV

An evidence-driven weekly planning Skill that limits preparation and prompts early external validation.

Reproducible pipeline for development and external validation of an early distant recurrence model in stage III colon cancer.

Heart disease prediction tested at a hospital the model never saw: leave-one-site-out AUROC, calibration, and conformal prediction sets on the 4-site UCI data

Analysis code and de-identified local validation cohort for a GCS-free long-term mortality model in critically ill patients with acute brain injury

Benchmark honesto de descriptores de imagen para clasificar lesiones cutaneas dermatoscopicas (ISIC): 390 combinaciones descriptor x algoritmo, control del efecto de lote y validacion externa entre instituciones. AUC 0.741 fuera del archivo de entrenamiento.