Your own 🤖 Doctor
-
Updated
Apr 28, 2023 - Python
Your own 🤖 Doctor
Machine-learning model that predicts hospital readmission risk from structured clinical data (Python, Jupyter, Docker).
흉부 X선과 EHR 로 30일 재입원을 예측하고, 누수, 보정, 서브그룹, 결측 모달리티로 성능 숫자를 검증함
EHR-style readmission and care-quality analytics project with Python, SQL, validation tests, and Power BI
Privacy-conscious Bronze-to-Gold healthcare data pipeline for patient identity resolution and disclosure-controlled readmission analytics.
30-day hospital readmission risk on 99,343 real inpatient encounters. Why AUC 0.66 is still useful, why balanced class weights break calibration, and what a care-management team actually gets.
Readmission risk judged at a fixed call capacity: on 101,766 UCI encounters split by patient, a 10% LightGBM call list reaches 25.0% of 30-day readmissions vs 20.5% for prior stays and 15.8% for LACE.
30-day hospital readmission prediction on 100K diabetic encounters: patient-level validation, logistic regression vs random forest, Tableau extracts.
To associate your repository with the readmission topic, visit your repo's landing page and select "manage topics."