Explainable Machine Learning for Classifying Iron–Folic Acid Supplementation Adherence Among Pregnant Women Attending Antenatal Care in Bosaso, Somalia: Dataset and Supporting Materials
Year: 2026
Type: dataset
Abstract: This dataset and supporting materials accompany a facility-based cross-sectional study of iron–folic acid supplementation (IFAS) adherence among 400 pregnant women attending antenatal care at eight health facilities in Bosaso, Somalia. The record contains the de-identified study dataset, source questionnaire, revised data dictionary, reproducible analysis scripts, model-performance outputs, calibration results, sensitivity-analysis outputs, and figures supporting the machine-learning analyses. This version was updated in response to peer review. The raw dataset and source questionnaire remain unchanged. The revised analytical materials correct categorical-variable encoding, use repeated stratified 5-fold cross-validation with 5 repeats for primary internal validation, provide bootstrap 95% confidence intervals for held-out performance, assess model calibration, include a sensitivity analysis using 7 tablets/week as a daily-use outcome, and use non-SMOTE Random Forest SHAP analysis for model interpretation. Previous integer-coded and SMOTE-based analytical outputs are superseded by the revised files in this version. The analyses remain exploratory and are intended for research and reproducibility purposes. The models should not be interpreted as causal or as ready for clinical deployment without external prospective validation.
Institutions: CITYCOT University, Amoud University
Language: English
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7203603168