jomarie04/hospital-readmission-predictor

---

model_name: hospital-readmission-predictor license: mit language: en tags: - tabular - classification - healthcare - patient-risk - synthetic-data metrics: - accuracy - f1 - roc_auc

Hospital Readmission Predictor

Model Overview

Hospital Readmission Predictor is a machine learning model trained to predict whether a patient will be readmitted to the hospital within 30 days.
The dataset is synthetic, based on typical electronic health record features.
Intended for research, ML demos, and education.

---

Model Details

• Model Name: hospital-readmission-predictor
• Model Type: Binary Classification
• Framework: scikit-learn / XGBoost
• Input Format: CSV / Tabular
• Output: Readmission risk (Yes / No)

---

Training Data

Synthetic patient dataset with features:

Feature	Type	Description
patient_id	String	Unique patient identifier
age	Integer	Patient age
gender	Categorical	Male / Female
diagnosis_code	Categorical	Primary diagnosis code
comorbidities	Integer	Number of comorbid conditions
length_of_stay	Integer	Hospital stay length in days
previous_admissions	Integer	Number of prior admissions
readmitted_30_days	Binary	Target label

---

Intended Use

✅ Research and ML tutorials
✅ Healthcare model demos (synthetic)
❌ Actual clinical decision-making

---

Evaluation Results

Metric	Score
Accuracy	0.84
F1-score	0.81
ROC-AUC	0.87

---

Limitations

• Synthetic patient data
• Limited diversity and complexity
• Not representative of real hospital data

---

Ethical Considerations

No real patient data is used.
Misuse for clinical purposes could risk patient safety.

---

License

This model is released under the MIT License.