DistanceModel: Create a DistanceModel object describing the network...

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

Description

Create a DistanceModel object describing the network structure of the population under study.

Usage

DistanceModel(
  distanceList,
  scaleMode = c("none", "rowscale", "invsqrt"),
  priorAlpha = 1,
  priorBeta = 1
)

Arguments

distanceList	a list of square, symmetric distance matrices.
scaleMode	an optional argument specifying the type of preprocessing needed.
priorAlpha	the first shape parameter for the beta distributed autocorrelation terms
priorBeta	the second shape parameter for the beta distributed autocorrelation terms

Details

In stochastic spatial SEIR models as specified in Brown et al. 2015, populations are divided into homogeneous groups, or locations, with heterogeneous mixing between groups. This is accomplished using a distance matrix parameterization, in which some number of square, symmetric distance matrices are constructed, each of which receives a spatial autocorrelation parameter.

Care must be taken to specify reasonable prior parameters for such terms, as well as in the construction and scaling of the distance matrices; it is certainly possible to construct an overspecified model, and to correspondingly bias inference about other important exposure process terms.

Value

an object of type DistanceModel

Examples

distanceModel <- DistanceModel(list(1-diag(4)))

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