fitMulti: Driver Routine - Coupled Spatial Model

In predsci/DICE: DICE: Dynamics of Interacting Community Epidemics

Description

A spatially coupled MCMC fit of the model data. The code first fits the model data directly and then fits it as a weighted sum of the coupled fit level data. This fit uses a coupling matrix to describe the interaction between different spatial regions and it generates all the fit level profiles at once and minimizes their weighted likelihood with the weights given by the relative population of each region. The data can be either cdc or gft data, and the model/fit data should have different spatial scales. For example in the case of cdc data: the model is national and the fit are the ten HHS regions. Or the model can be an HHS region and the fit fit data is state level data.

Usage

fitMulti(mydata = NULL, all_years_epi = NULL, opt.list = NULL,
  run.list = NULL, ireal = 1, iseed = NULL)

Arguments

mydata	A complex list with all available data for a given disease season model/fit spatial levels and data type
opt.list	A Logical list with TRUE or FALSE values for all the parameters supported by DICE. These values are based on the user's chosen compartmental model (SIR/SEIR) and the force of infection.
run.list	A list with parameters needed for the MCMC procedure
ireal	Integer - the MCMC chain number. Default is 1.
iseed	An integer used to set the random-number-generator seed. When not specified, the seed is generated randomly. Setting the seed to a known integer allows an MCMC chain to be reproducible.

Value

A list with the following arguments:

model_rtn

The best result of the MCMC procedure for the model data

model_profile

Randomly selected results from the MCMC procedure of directly fitting the model data

tab.model

The MCMC history of the direct fit to the model data

rtn

The best result for indirectly fitting the model data using the fit regions

profile

Randomly selected results from the MCMC procedure of indirectly fitting the model data using the fit regions

tab

The MCMC history of indirectly fitting the model data using the fit regions

Examples

fitMulti{mydata = mydata, opt.list = opt.list, run.list = run.list, ireal = ireal, iseed = iseed}

predsci/DICE documentation built on Aug. 9, 2019, 9:41 a.m.