DataModel: Create a DataModel object, describing the relationship of...

In grantbrown/ABSEIR: Posterior Approximation to Fully Bayesian Spatial SEIR Models Using Multithreaded C++

Description

Create a DataModel object, describing the relationship of your data to the SEIR model

Usage

DataModel(
  Y,
  type = c("identity", "overdispersion"),
  compartment = c("I_star", "R_star", "I"),
  cumulative = FALSE,
  params = list()
)

Arguments

Y	A matrix with T rows and n columns, for time points and spatial locations respectively.
type	A string equal to "identity", "overdispersion", or "fractional"
compartment	A string equal to "I_star" or "R_star"
cumulative	A logical value indicating whether the data is reported on a cumulative scale
params	A list of optional parameters

Details

A fundamental task in building hierarchical models is to describe the way in which the observed data relates to the underlying model. Here, we assume that the matrix of observed values, Y, is related to one of two epidemic compartments: I_star or R_star. In the notation of Brown, Porter, and Oleson 2015, these correspond to the $I^*$ and $R^*$ compartments, which catalog the newly infectious and recovered individuals, respectively. For a discussion of these definitions, in addition to a more detailed description of the overall spatial SEIR framework, please refer to that work. Depending on the data model type, additional parameters may be required:

• phi: An overdispersion parameter, required for the 'overdispersion' data model type.

• report_fractionThe (scalar) estimated reporting fraction for the epidemic. Required for the 'fractional' data model.

• report_fraction_essThe effective sample size associated with the required reporting fraction in the 'fractional' data model.

Value

an object of type DataModel

Examples

dataModel = DataModel(rpois(100, 10)) 

grantbrown/ABSEIR documentation built on Oct. 14, 2021, 2:32 p.m.