Bayesian modelling for individual-level infectious disease models

由于全球新冠肺炎大流行，传染病模型从未像现在这样重要。疾病传播的数学模型在世界各地被广泛使用，以确定防止感染传播的最有效干预措施。然而，为了提供对疾病动力学的有效洞察，数学模型必须以有效的方式与可用的数据相匹配。这是具有挑战性的，因为(I)复杂的计算方法是唯一被证实的拟合此类模型的方法，(Ii)传播过程在很大程度上是未被观察到的。相反，模型适用于传播的间接观察，如阳性检测、住院和死亡。该项目的目的将是开发将有效的传染病模型与数据相匹配的方法。具体地说，建立了个体水平协变量值对易感性、传染性和疫苗保护水平等参数的影响的模型。潜在的应用将是改进传染病的建模，特别是新出现的疾病，帮助预防和应对大流行和未来的大流行。由于这种模型产生的参数数量很多，因此需要先进的计算技术。这里特别关注的是贝叶斯环境中的传染病模型，以及使用蒙特卡洛方法来帮助得出推断。将使用各种模拟和真实世界的数据集来建立和开发模型。

Modelling infectious diseases has never been more important due to the global COVID-19 pandemic. Mathematical models of disease transmission are in heavy use across the world to identify the most effective interventions to prevent the spread of infection. However, to provide effective insight into the disease dynamics, mathematical models must be fit to the available data in an effective way. This is challenging because (i) complex, computational methods are the only proven way of fitting such models and (ii) the transmission process is largely unobserved. Models are instead fitted to indirect observations of transmission, such as positive tests, hospitalisations, and deaths. The aim of this project will be to develop methods for fitting efficient infectious disease models to data. Specifically, models which build in the effect of individual-level covariate values on parameters such as susceptibility, infectiousness and vaccine protection levels. The potential applications will be to improve modelling of infectious diseases, especially emerging diseases, helping to prevent and deal with the pandemic and future pandemics. Due to the high number of parameters that such models yield, advanced computational techniques will be required. Of particular focus here will be infectious disease models in a Bayesian setting and the use of Monte Carlo methods to help draw inference. A variety of simulated and real-world datasets will be used to build and develop models.