Abstract:
|
Compartment models (CM) and agent-based models (AM) are two common classes of models used to predict the spread of infectious disease in a region and to infer details about the disease such as the rate of infection. Both classes describe how individuals travel from susceptible states to infectious states (compartments). CMs rely on assumptions of homogeneity, i.e. individuals in the population act in the same manner. On the other hand, AMs allow for heterogeneity of individuals (agents). Individuals in CMs travel from one compartment to another based on ordinary differential equations whereas AMs are based on simulation. Despite these differences, in Gallagher and Eddy 2018, we describe conditions in which these two classes are equal in joint distribution, allowing for hybrid models. Here, we demonstrate this hybrid model through the program "Compartment and agent-based temporal models and analysis" (Catalyst). We experiment using independent and dependent agents that live in different households and attend different activities. We analyze the distances between the CMs and AMs and create a threshold to switch between models. Ultimately, we will create a seamless hybrid model.
|