Scholarlink

The recent global health crisis stemming from the resurgence of monkeypox outbreaks emphasizes the imperative for transparent and dependable predictive tools to guide public health interventions. The recent global health crisis stemming from the resurgence of monkeypox outbreaks emphasizes the imperative for transparent and dependable predictive tools to guide public health interventions. In response to this exigency, this study proposes a novel framework that employs Bayesian network modeling for the assessment of monkeypox transmission likelihood. This approach formulates a robust probabilistic model by synthesizing essential determinants including epidemiological patterns, clinical symptoms, and socioeconomic factors pertinent to disease spread. Trained and validated utilizing data from recent outbreaks, achieving an AUC of 0.88 in cross-validation. Importantly, the intrinsic interpretability of the Bayesian network approach offers certain advantages. Unlike opaque machine learning models, it clearly shows how specific key risk predictors such as close physical contact, presence of skin lesions and travel history to endemic regions contribute to transmission probability. This clarity enhances its applicability in both clinical risk assessment and the design of public health strategies. The insights generated by the model can inform targeted interventions, optimize resource allocation, and enhance effective risk communication. Future research endeavors should focus on integrating real-time surveillance data and augmenting the model to capture the dynamic nature of disease transmission over time.

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

Predictive Modeling and Risk Assessment of Monkeypox Transmission Using Bayesian Networks: An Interpretable Machine Learning Approach

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Zvinodashe Revesai,
  • Kudakwashe Maguraushe,
  • Belinda Ndlovu,
  • Sibusisiwe Dube

摘要

The recent global health crisis stemming from the resurgence of monkeypox outbreaks emphasizes the imperative for transparent and dependable predictive tools to guide public health interventions. The recent global health crisis stemming from the resurgence of monkeypox outbreaks emphasizes the imperative for transparent and dependable predictive tools to guide public health interventions. In response to this exigency, this study proposes a novel framework that employs Bayesian network modeling for the assessment of monkeypox transmission likelihood. This approach formulates a robust probabilistic model by synthesizing essential determinants including epidemiological patterns, clinical symptoms, and socioeconomic factors pertinent to disease spread. Trained and validated utilizing data from recent outbreaks, achieving an AUC of 0.88 in cross-validation. Importantly, the intrinsic interpretability of the Bayesian network approach offers certain advantages. Unlike opaque machine learning models, it clearly shows how specific key risk predictors such as close physical contact, presence of skin lesions and travel history to endemic regions contribute to transmission probability. This clarity enhances its applicability in both clinical risk assessment and the design of public health strategies. The insights generated by the model can inform targeted interventions, optimize resource allocation, and enhance effective risk communication. Future research endeavors should focus on integrating real-time surveillance data and augmenting the model to capture the dynamic nature of disease transmission over time.