Almutiry Waleed, Deardon Rob
Mathematics, Arts and Science College in Ar Rass, Qassim University, Buraidah, Saudi Arabia.
Production Animal Health, University of Calgary, Calgary, Alberta, Canada.
Stat Commun Infect Dis. 2021 Jan 8;13(1):20190012. doi: 10.1515/scid-2019-0012. eCollection 2021 Jan 1.
Infectious disease transmission between individuals in a heterogeneous population is often best modelled through a contact network. This contact network can be spatial in nature, with connections between individuals closer in space being more likely. However, contact network data are often unobserved. Here, we consider the fit of an individual level model containing a spatially-based contact network that is either entirely, or partially, unobserved within a Bayesian framework, using data augmented Markov chain Monte Carlo (MCMC). We also incorporate the uncertainty about event history in the disease data. We also examine the performance of the data augmented MCMC analysis in the presence or absence of contact network observational models based upon either knowledge about the degree distribution or the total number of connections in the network. We find that the latter tend to provide better estimates of the model parameters and the underlying contact network.
在异质人群中，个体间的传染病传播通常通过接触网络进行最佳建模。这种接触网络本质上可能是空间性的，空间距离较近的个体之间建立联系的可能性更大。然而，接触网络数据往往难以观测到。在此，我们考虑一个个体层面的模型，该模型包含一个基于空间的接触网络，在贝叶斯框架内，这个网络要么完全未被观测到，要么部分未被观测到，我们使用数据增强马尔可夫链蒙特卡罗（MCMC）方法进行分析。我们还将疾病数据中事件历史的不确定性纳入其中。我们还研究了在存在或不存在基于网络度分布知识或网络连接总数的接触网络观测模型的情况下，数据增强MCMC分析的性能。我们发现，后者往往能对模型参数和潜在接触网络提供更好的估计。
