Quick Simulation from an Endemic-Only twinstim

In endemic-only twinstim models, the conditional intensity is a piecewise constant function independent from the history of the process. This allows for a much more efficient simulation algorithm than via Ogata's modified thinning as in the general simulate.twinstim method.

Usage

simEndemicEvents(object, tiles)

Arguments

object

an object of class "twinstim" (with the model component retained; otherwise try object <- update(object, model = TRUE)).

tiles

an object inheriting from "SpatialPolygons", which represents the tiles of the original data's stgrid (see, e.g., levels(environment(object)$gridTiles)).

Value

a "SpatialPointsDataFrame"

Author

Sebastian Meyer

See also

the general simulation method simulate.twinstim

Examples

data("imdepi", "imdepifit")
load(system.file("shapes", "districtsD.RData", package="surveillance"))

## Fit an endemic-only twinstim()
m_noepi <- update(imdepifit, epidemic = ~0, siaf = NULL, model = TRUE,
                  T = 120)  # using a restricted time range, for speed

## Simulate events from the above endemic model
set.seed(1)
s1 <- simEndemicEvents(m_noepi, tiles = districtsD)
class(s1)  # just a "SpatialPointsDataFrame"
summary(s1@data)
plot(imdepi$W, lwd = 2, asp = 1)  
plot(s1, col = s1$type, cex = 0.5, add = TRUE)

## Compare with the generic simulation method (slower)
s0 <- simulate(m_noepi, seed = 1, data = imdepi, tiles = districtsD)
class(s0)  # gives a full "simEpidataCS" with several methods applicable
methods(class = "epidataCS")
plot(s0, "time")
plot(s0, "space", points.args = list(pch = 3), lwd = 2)