Scholarlink

This chapter advocates for the use of Markov Chain Monte Carlo (MCMC) and neural networks methods for uncertainty quantification in epidemiological inverse problems. Accurate calibration of epidemiological model parameters from contagion data enhances the understanding of the disease dynamics and supports the formulation of scenarios for informed decision-making. These inverse problems have been addressed within the uncertainty quantification framework to characterize the statistical uncertainty of the results. Meanwhile, machine learning methods, particularly artificial neural networks, have been widely adopted across disciplines due to their remarkable capacity to capture complex patterns and uncover underlying relationships in data. Consequently, artificial neural networks are increasingly employed for various purposes, including epidemiological modeling. This chapter presents a review of MCMC methods and introduces neural network implementations, in the context of a classical epidemiological model, describing how they are used for uncertainty quantification purposes. Specific aspects of each method are discussed, along with remarks based on a real-world COVID-19 contagion data scenario.

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

An Introduction to Neural Networks for Uncertainty Quantification in Epidemiological Models

  • 动态工具书词条

  • 发表日期：

    2025年

查看全文
  • Abel Palafox González,
  • L. Leticia Ramírez-Ramírez,
  • Román Zúñiga-Macías,
  • Sergio Barajas-Oviedo,
  • Ulises Uriostegui-Legorreta

摘要

This chapter advocates for the use of Markov Chain Monte Carlo (MCMC) and neural networks methods for uncertainty quantification in epidemiological inverse problems. Accurate calibration of epidemiological model parameters from contagion data enhances the understanding of the disease dynamics and supports the formulation of scenarios for informed decision-making. These inverse problems have been addressed within the uncertainty quantification framework to characterize the statistical uncertainty of the results. Meanwhile, machine learning methods, particularly artificial neural networks, have been widely adopted across disciplines due to their remarkable capacity to capture complex patterns and uncover underlying relationships in data. Consequently, artificial neural networks are increasingly employed for various purposes, including epidemiological modeling. This chapter presents a review of MCMC methods and introduces neural network implementations, in the context of a classical epidemiological model, describing how they are used for uncertainty quantification purposes. Specific aspects of each method are discussed, along with remarks based on a real-world COVID-19 contagion data scenario.