--- license: cc-by-4.0 task_categories: - tabular-classification language: - en tags: - healthcare - maternal-health - bangladesh - fairness - explainable-ai - clinical-validation size_categories: - n<10K --- # Bangladesh Maternal Health — Research Data Research outputs from the maternal health XAI study by [Farjana Yesmin](https://farjana-yesmin.github.io): > Yesmin, F., Shirmin, N. & Bristy, S.S. (2026). > *Explainable AI for Maternal Health Risk Prediction in Bangladesh: > A Hybrid Fuzzy-XGBoost Framework with Clinician Validation.* > ICAIHE 2026, Waseda University, Tokyo. > [Preprint](https://www.researchsquare.com/article/rs-8584734/v1) Part of the [FairHealth](https://github.com/Farjana-Yesmin/fairhealth) library — `pip install fairhealth` ## What Is Included | File | Description | |---|---| | `final_corrected_results.csv` | Final model evaluation results | | `model_comparison_results.csv` | 7 models compared (accuracy, F1, AUC) | | `results_summary.csv` | Summary statistics across experiments | | `xai_comparison_results.csv` | SHAP vs Fuzzy vs Hybrid explanation comparison | | `Explainability Evaluation...csv` | Clinician validation survey (n=14) | | `final_fuzzy_xgboost_model.json` | Trained Fuzzy-XGBoost model weights | ## Key Results | Model | Accuracy | F1 | ROC-AUC | |---|---|---|---| | **Hybrid Fuzzy-XGBoost (ours)** | **88.67%** | **0.8869** | **0.9703** | | Gradient Boosting | 86.21% | 0.8619 | — | | Random Forest | 85.22% | 0.8521 | — | Clinician validation (N=14): - **71.4%** preferred hybrid Fuzzy+SHAP explanation - **54.8%** would trust in clinical practice - Top barrier: internet connectivity (50%) ## Data Sources - UCI Maternal Health Risk Dataset (base clinical data) - DGHS Dashboard 2024 — Bangladesh regional healthcare indicators - UNFPA MPDSR 2023 — Maternal mortality ratios by division - WHO ANC guidelines, ADA 2023, ACC/AHA 2017 (fuzzy rules) ## Usage ```python from fairhealth.explain.fuzzy import get_fired_rules rules = get_fired_rules(age=42, sbp=145, bs=12, hr=88) for r in rules: print(f"Rule {r['id']}: {r['condition']} → {r['outcome']}") ``` ## Citation ```bibtex @dataset{fairhealth_maternal_2026, author = {Yesmin, Farjana}, title = {Bangladesh Maternal Health Research Dataset}, year = {2026}, publisher = {Hugging Face}, doi = {10.57967/hf/8801}, url = {https://huggingface.co/datasets/fairhealth/bangladesh-maternal-health} } ``` Also cite: ```bibtex @inproceedings{yesmin2026icaihe, author = {Yesmin, Farjana and Shirmin, Nusrat and Bristy, Suraiya Shabnam}, title = {Explainable AI for Maternal Health Risk Prediction in Bangladesh}, note = {ICAIHE 2026, Waseda University, Tokyo}, year = {2026} } ```