Scholarlink

<p>Under-five mortality remains a critical indicator of child health and social well-being, reflecting disparities in healthcare access, socioeconomic conditions, and environmental factors. This study applies a geographically weighted regression model (GWRM) to analyze the spatial heterogeneity of under-five mortality across 18 provinces in Iraq, utilizing data from 2022. Unlike traditional global regression models, GWRM model accounts for local variation in the relationships between mortality and its determinants, capturing region-specific risk patterns. Our study compares four kernel functions for bandwidth optimization, identifying the Bi-square kernel as optimal with the highest explanatory power (pseudo-R² = 0.9309) and lowest Bayesian Information Criterion (BIC = 103.661). Results reveal significant spatial variation in the influence of variables including rural population percentage, fertility rates, growth rate, dependency ratio, median age, and sex ratio on under-five mortality. The GWRM model explains 93.09% of variability in mortality rates, outperforming the global model. Residual analysis indicates localized prediction errors, highlighting areas for further investigation. These spatially explicit findings underscore the importance of geographically targeted public health interventions to address child mortality disparities in Iraq effectively.</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

Geospatial modeling of under-five mortality in Iraq based on geographic weighted regression model

  • 期刊论文

  • 发表日期：

    2025年12月19日

查看全文
  • Omar Fawzi Salih Al-Rawi,
  • Zakariya Yahya Algamal

摘要

Under-five mortality remains a critical indicator of child health and social well-being, reflecting disparities in healthcare access, socioeconomic conditions, and environmental factors. This study applies a geographically weighted regression model (GWRM) to analyze the spatial heterogeneity of under-five mortality across 18 provinces in Iraq, utilizing data from 2022. Unlike traditional global regression models, GWRM model accounts for local variation in the relationships between mortality and its determinants, capturing region-specific risk patterns. Our study compares four kernel functions for bandwidth optimization, identifying the Bi-square kernel as optimal with the highest explanatory power (pseudo-R² = 0.9309) and lowest Bayesian Information Criterion (BIC = 103.661). Results reveal significant spatial variation in the influence of variables including rural population percentage, fertility rates, growth rate, dependency ratio, median age, and sex ratio on under-five mortality. The GWRM model explains 93.09% of variability in mortality rates, outperforming the global model. Residual analysis indicates localized prediction errors, highlighting areas for further investigation. These spatially explicit findings underscore the importance of geographically targeted public health interventions to address child mortality disparities in Iraq effectively.