Project Overview
Hospital readmissions within 30 days are costly and often preventable. Our model predicts which patients are at high risk, allowing for targeted interventions.
Using machine learning on the UCI Diabetes dataset from 130 US hospitals (1999-2008), we've built a model that identifies key readmission risk factors.
Our model can help hospitals reduce readmission rates by 20-30%, improve patient outcomes, and save millions in healthcare costs.
Key Features
Our comprehensive solution provides healthcare providers with powerful tools to identify and manage readmission risks.
Visualize readmission patterns, risk factors, and model performance metrics through interactive charts and graphs.
Input patient data to receive an instant readmission risk assessment with personalized intervention recommendations.
Model Performance
Our model achieves high accuracy in predicting 30-day readmissions while providing explainable results.
Accuracy
Precision
Recall
F1 Score
About the Dataset
This project uses the UCI Diabetes 130-US hospitals dataset, covering 10 years (1999-2008) of clinical care.
Dataset Features:
- Patient demographics (age, gender, race)
- Admission type and source
- Length of hospital stay
- Primary and secondary diagnoses
- Medication data
- Lab test results
- Readmission outcomes
Dataset Statistics:
- 101,766 hospital admissions
- 71,518 unique patients
- 130 US hospitals
- 50+ features per patient
- Balanced class distribution