Integrating Feature Selection, Machine Learning, and SHAP Explainability to Predict Severe Acute Pancreatitis.
Source check: Source checked. Not independent scientific confirmation.
Finding: A retrospective single-center cohort of 743 adults (67 severe cases) was used to predict severe acute pancreatitis from ED-arrival variables; the top RFE-RF+kNN pipeline had test-set AUROC 0.826.
Technology: Six feature selectors were combined with six ML classifiers; SHAP ranked variable contributions.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Six feature selectors were combined with six ML classifiers; SHAP ranked variable contributions.
- Test and result
- The best RFE-RF-feature+kNN pipeline predicted Revised Atlanta severe disease (AUROC 0.826; 95% CI 0.686–0.965) in the study cohort.
- Limitation / still unknown
- A retrospective single-center cohort of 743 adults (67 severe cases) was used to predict severe acute pancreatitis from ED-arrival variables; the top RFE-RF+kNN pipeline had test-set AUROC 0.826. A random 70/30 split was used, with five-fold training CV and performance on the independent internal 30% test set; bootstrap CIs and calibration were reported. Retrospective single-center results need external validation and do not establish improved outcomes.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.mdpi.com/2075-4418/15/19/2473
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.