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.
-
Updated
Aug 14, 2026 - Python
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)
Does a learned image-enhancement front-end help a retinal vessel segmenter on data it has never seen? A controlled study with a fixed backbone and pre-registered rules: it gains on leave-one-dataset-out benchmarks and loses on an external clinical dataset.
Heart disease prediction V2: leakage audit, leave-one-hospital-out validation, calibration and modern tabular models
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
Predictions-only calibration validation for binary clinical prediction models: cluster-aware, cross-fitted, report-emitting. Python.
Research software for consensus clustering, resampling stability, and external-label evaluation.
Interpretable TCGA-BRCA RNA-seq breast cancer subtype classification with ML, SHAP, pathway enrichment, and METABRIC external validation.
Pipeline de recherche sur le fond d'œil (RETFound, RETFound-Green) : classification multi-pathologies, validation externe sur APTOS et EyePACS, localisation des lésions sur IDRiD. Pas un dispositif médical.
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.
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.
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
512-patient MedRAX CXR classifier external-validation pilot on TAIX-Ray, with a tested 16-bit input compatibility audit; no raw data published.
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 and de-identified local validation cohort for a GCS-free long-term mortality model in critically ill patients with acute brain injury
Reproducible datasets, computational workflows, and model implementations for externally validated tabular foundation models and explainable concrete compressive strength prediction.
To associate your repository with the external-validation topic, visit your repo's landing page and select "manage topics."