Threshold Optimized Random Forest and XGBoost with SHAP for Low Birth Length Risk Prediction

Rian andi saputra; Albert Yakobus Chandra

doi:10.55537/cosie.v5i3.1825

Authors

Rian andi saputra Mercu Buana University Yogyakarta
Albert Yakobus Chandra Mercu Buana University Yogyakarta

DOI:

https://doi.org/10.55537/cosie.v5i3.1825

Keywords:

Random Forest, XGBoost, Birth Length Risk, Threshold Optimization, SHAP

Abstract

Stunting is a chronic nutritional disorder that impairs children's cognitive development and longterm productivity, making early risk identification essential. However, previous studies have mainly focused on stunting classification or conventional machine learning models, with limited attention to integrating probability threshold optimization and explainable artificial intelligence (XAI) for low birth length risk prediction. This study proposes a predictive framework using Random Forest (RF) and XGBoost based on 249,626 valid records from the 2024 Indonesian Nutritional Status Survey (SSGI), where low birth length was defined as a birth length below 48 cm. The models were optimized through hyperparameter tuning, probability threshold optimization, Stratified 5-Fold Cross Validation, and SHAP analysis. Evaluation on 10,000 testing records showed that Random Forest accuracy increased from 76.76% to 78.58%, while XGBoost achieved the highest accuracy of 84.07% at a threshold of 0.79, although recall decreased to 27%. SHAP identified birth weight as the most influential predictor. These findings demonstrate that threshold optimization improves predictive accuracy, while SHAP enhances model interpretability to support early screening and clinical decision-making for low birth length risk

Downloads

Download data is not yet available.

References

[1] Y. Xiong, X. Hu, J. Cao, L. Shang, and B. Niu, “A predictive model for stunting among children under the age of three,” Front. Pediatr., vol. 12, Sep. 2024, doi: 10.3389/fped.2024.1441714.

[2] G. Nduwayezu et al., “Spatial Machine Learning for Exploring the Variability in Low Height‐For‐Age From Socioeconomic, Agroecological, and Climate Features in the Northern Province of Rwanda,” Geohealth, vol. 8, no. 9, Sep. 2024, doi: 10.1029/2024GH001027.

[3] A. Ranjbar, F. Montazeri, M. V. Farashah, V. Mehrnoush, F. Darsareh, and N. Roozbeh, “Machine learning-based approach for predicting low birth weight,” BMC Pregnancy Childbirth, vol. 23, no. 1, p. 803, Nov. 2023, doi: 10.1186/s12884-023-06128-w.

[4] A. Naik, G. G. Tejani, and S. J. Mousavirad, “SGO enhanced random forest and extreme gradient boosting framework for heart disease prediction,” Sci. Rep., vol. 15, no. 1, p. 18145, May 2025, doi: 10.1038/s41598-025-02525-7.

[5] C. Kalinda et al., “Leveraging multisectoral approach to understand the determinants of childhood stunting in Rwanda: a systematic review and meta-analysis,” Syst. Rev., vol. 13, no. 1, p. 16, Jan. 2024, doi: 10.1186/s13643-023-02438-4.

[6] Z. Fan et al., “XGBoost-SHAP-based interpretable diagnostic framework for knee osteoarthritis: a population-based retrospective cohort study,” Arthritis Res. Ther., vol. 26, no. 1, Dec. 2024, doi: 10.1186/s13075-024-03450-2.

[7] H. Shen, H. Zhao, and Y. Jiang, “Machine Learning Algorithms for Predicting Stunting among Under-Five Children in Papua New Guinea,” Children, vol. 10, no. 10, p. 1638, Sep. 2023, doi: 10.3390/children10101638.

[8] T. K. Wudu, A. A. Endalew, and A. A. Dires, “Explainable hybrid machine learning model for predicting stunting and identifying key risk factors among Ethiopian children under five,” Sci. Rep., vol. 16, no. 1, Dec. 2026, doi: 10.1038/s41598-026-46417-w.

[9] J. Munyemana, I. H. Kabano, B. Uzayisenga, A. R. Cyamweshi, E. Ndagijimana, and E. Kubana, “The role of national nutrition programs on stunting reduction in Rwanda using machine learning classifiers: a retrospective study,” BMC Nutr., vol. 10, no. 1, p. 98, Jul. 2024, doi: 10.1186/s40795-024-00903-4.

[10] N. El Furqany, M. Subianto, and A. Rusyana, “Hybrid Ensemble Learning with SMOTEENN and Soft Voting for Stunting Risk Prediction: A SHAP-Based Explainable Approach,” Journal of Applied Data Sciences, vol. 6, no. 4, pp. 2989–3004, Dec. 2025, doi: 10.47738/jads.v6i4.829.

[11] J. L. Leevy, J. M. Johnson, J. Hancock, and T. M. Khoshgoftaar, “Threshold optimization and random undersampling for imbalanced credit card data,” J. Big Data, vol. 10, no. 1, p. 58, May 2023, doi: 10.1186/s40537-023-00738-z.

[12] V. Vimbi, N. Shaffi, and M. Mahmud, “Interpreting artificial intelligence models: a systematic review on the application of LIME and SHAP in Alzheimer’s disease detection,” Brain Inform., vol. 11, no. 1, p. 10, Dec. 2024, doi: 10.1186/s40708-024-00222-1.

[13] B. Paneru, B. Paneru, S. C. Sapkota, and R. Poudyal, “Enhancing healthcare with AI: Sustainable AI and IoT-Powered ecosystem for patient aid and interpretability analysis using SHAP,” Measurement: Sensors, vol. 36, p. 101305, Dec. 2024, doi: 10.1016/j.measen.2024.101305.

[14] L. Malinverno et al., “A historical perspective of biomedical explainable AI research,” Patterns, vol. 4, no. 9, p. 100830, Sep. 2023, doi: 10.1016/j.patter.2023.100830.

[15] G. M. Setegn and B. E. Dejene, “Improving machine learning models through explainable AI for predicting the level of dietary diversity among Ethiopian preschool children,” Ital. J. Pediatr., vol. 51, no. 1, p. 91, Mar. 2025, doi: 10.1186/s13052-025-01892-1.

[16] M. Sinaga, F. Fujiati, and D. Halawa, “Designing a Stunting Prediction Model Using Machine Learning to Support SDGs Achievement in Indonesia,” sinkron, vol. 9, no. 4, pp. 2070–2079, Oct. 2025, doi: 10.33395/sinkron.v9i4.15296.

[17] T. Chen and C. Guestrin, “XGBoost,” in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, New York, NY, USA: ACM, Aug. 2016, pp. 785–794. doi: 10.1145/2939672.2939785.

[18] T. O. Omotehinwa, D. O. Oyewola, and E. G. Dada, “A Light Gradient-Boosting Machine algorithm with Tree-Structured Parzen Estimator for breast cancer diagnosis,” Healthcare Analytics, vol. 4, p. 100218, Dec. 2023, doi: 10.1016/j.health.2023.100218.

[19] Md. J. Raihan, Md. A.-M. Khan, S.-H. Kee, and A.-A. Nahid, “Detection of the chronic kidney disease using XGBoost classifier and explaining the influence of the attributes on the model using SHAP,” Sci. Rep., vol. 13, no. 1, p. 6263, Apr. 2023, doi: 10.1038/s41598-023-33525-0.

[20] Md. Z. Islam, M. R. K. Chowdhury, M. Kader, B. Billah, Md. S. Islam, and M. Rashid, “Determinants of low birth weight and its effect on childhood health and nutritional outcomes in Bangladesh,” J. Health Popul. Nutr., vol. 43, no. 1, p. 64, May 2024, doi: 10.1186/s41043-024-00565-9.

[21] S. M. Lundberg et al., “From local explanations to global understanding with explainable AI for trees,” Nat. Mach. Intell., vol. 2, no. 1, pp. 56–67, Jan. 2020, doi: 10.1038/s42256-019-0138-9.