Beyond prediction: explainable machine learning for ferritin estimation from complete blood count.
Source check: Source checked. Not independent scientific confirmation.
Finding: In 12,450 adult laboratory records, five supervised models estimated low ferritin from complete blood count and demographic features. On the balanced 7,200-sample dataset, the artificial neural network had the best low-ferritin classification result (AUC 0.837; F1 0.760); regression had global R² 0.32, with stronger performance in lower ferritin ranges. The evaluation used one laboratory source and does not establish external screening validity.
Technology: Five supervised models, including logistic regression, XGBoost and an artificial neural network, classified ferritin below 30 ng/mL and also estimated continuous ferritin from CBC and demographic features; SHAP summarized feature contributions.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Five supervised models, including logistic regression, XGBoost and an artificial neural network, classified ferritin below 30 ng/mL and also estimated continuous ferritin from CBC and demographic features; SHAP summarized feature contributions.
- Test and result
- The primary abstract and Results report analysis of 12,450 adult laboratory records, a balanced 7,200-sample dataset, five supervised algorithms, ANN low-ferritin classification AUC 0.837 and F1 0.760, and regression global R² 0.32. The input review notes cross-validation and held-out testing, with single-source records and moderate regression fit limiting screening interpretation.
- Limitation / still unknown
- All records came from a single laboratory source; no independent external validation or clinically validated screening use is established, and continuous ferritin regression had global R² 0.32.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC13439977/
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 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.