Scholarlink

<p>Infant mortality is a significant public health challenge in developing nations, particularly in India, where effective predictive models are critical for targeted interventions. This study utilizes the XGBoost algorithm, a robust machine learning technique, to predict infant mortality using socio-demographic, maternal, and child health variables from the National Family Health Survey-4 (2015–2016). The dataset, derived from a stratified multistage sampling method, includes comprehensive information on key factors influencing infant mortality. Key predictors, including maternal age, birth weight, and access to healthcare, were identified through feature importance analysis. The XGBoost model was trained and evaluated using a split dataset, demonstrating exceptional performance metrics: 97.6% accuracy, 96.8% precision, 97.2% recall, 96.9% F1-score, and an AUC-ROC of 0.974. The findings reveal that the XGBoost model outperforms traditional approaches, effectively addressing challenges such as data sparsity and class imbalance. This study highlights the utility of advanced machine learning methods in predicting infant mortality, providing actionable insights for healthcare professionals and policymakers.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Advancing public health initiatives: a comprehensive analysis of infant mortality predictors in India using the robust XGBoost algorithm

  • 期刊论文

  • 发表日期：

    2025年3月4日

查看全文
  • Indu Verma,
  • Sanjeev Kumar Prasad

摘要

Infant mortality is a significant public health challenge in developing nations, particularly in India, where effective predictive models are critical for targeted interventions. This study utilizes the XGBoost algorithm, a robust machine learning technique, to predict infant mortality using socio-demographic, maternal, and child health variables from the National Family Health Survey-4 (2015–2016). The dataset, derived from a stratified multistage sampling method, includes comprehensive information on key factors influencing infant mortality. Key predictors, including maternal age, birth weight, and access to healthcare, were identified through feature importance analysis. The XGBoost model was trained and evaluated using a split dataset, demonstrating exceptional performance metrics: 97.6% accuracy, 96.8% precision, 97.2% recall, 96.9% F1-score, and an AUC-ROC of 0.974. The findings reveal that the XGBoost model outperforms traditional approaches, effectively addressing challenges such as data sparsity and class imbalance. This study highlights the utility of advanced machine learning methods in predicting infant mortality, providing actionable insights for healthcare professionals and policymakers.