Views
No views yet
⚠️ Research prototype — NOT for clinical use. This model is trained on 117 visits from 30 patients. It informs clinical judgment; it does not replace it, and must never be the sole basis of a care decision. See the ethics statement.
Pipeline
(unified preprocessing + tuned Gradient Boosting) that estimates the risk of
hospitalisation within one year for elderly patients, from routine geriatric
data (gait speed, medications, frailty, cognition, falls, …).Note: this model embeds a custom preprocessing class from thepraeviuspackage, so the Hub's hosted inference widget cannot run it. Use the Space above, or load it locally with the snippet below.
| Metric | Value |
|---|---|
| Model | Gradient Boosting (tuned) |
| Honest AUC (nested cross-validation) | 0.739 ± 0.277 |
| Evaluation | patient-level nested_cv |
| Training data | 30 patients · 117 visits |
praevius package (loads this same model, bundled):pip install praevius # or: pip install "git+https://github.com/Zanarino/praevius.git"1import praevius
2models = praevius.load_pipelines()
3df = praevius.patient_to_dataframe({"age": 79, "gender": "F", "fried": "Pre-frail",
4 "gait_speed_4m": 0.75, "falls_one_year": 1})
5scores = praevius.score_patient(models, df)
6prob = scores["1year"]["Gradient Boosting"]
7band = praevius.risk_band(prob) # 'low' | 'moderate' | 'high'praevius package, so install it first (pip install praevius), then:1from huggingface_hub import hf_hub_download
2from skops.io import load, get_untrusted_types
3import praevius # registers praevius.preprocessing.ClinicalPreprocessor
4
5f = hf_hub_download("Zanarino/praevius-hospitalization-1year", "praevius_gb_tuned_1year.skops")
6# Review the types before trusting them (they are this project's classes):
7trusted = get_untrusted_types(file=f)
8pipeline = load(f, trusted=trusted)1trusted = [
2 "praevius.preprocessing.ClinicalPreprocessor",
3]praevius_gb_tuned_1year.skops — the champion 1-year pipeline (skops format).model_card.json — the full machine-readable card (all horizons, honest metrics, calibration decision).