Biomarker Performance — ROC·AUC·cutoff와 불확실성
What it does Binary reference label과 continuous biomarker score를 읽어 empirical ROC/AUC, seeded stratifiedbootstrap AUC 95% interval, Youden J 후보 cutoff, confusion matrix, sensitivity·specificity·PPV·NPV·accuracy와 Wilson interval을 계산합니다. Quick start 브라우저에서 바로 실행하거나 repository를 내려받아 python3 m http.server 8080으로 실행합니다. Inputs sample, truth, score 열을 가진 CSV. truth는 positive/negative, case/control, 1/0을 지원합니다. Outputs ROC plot, AUC interval, editable threshold metrics, sample classification과 provenance CSV. Scientific method Pairwiseconcordance empirical AUC, Youden J, Wilson score interval, fixedseed classstratified bootstrap을 사용합니다. 동일 데이터 cutoff 선택·평가의 optimism을 명시합니다. Privacy All calculations run locally. Do not upload confidential data to thirdparty services. Limitations This software is a calculation and exploratory analysis aid. It does not diagnose individuals, select a clinical cutoff, or replace independent validation, professional review, or regulatory decisions.