fitsir: fitting function

In bbolker/fitsir: SIR fitting tools

Description

fitting function

Usage

fitsir(data, start, dist = c("gaussian2", "gaussian", "poisson",
  "quasipoisson", "nbinom", "nbinom1"), type = c("prevalence",
  "incidence", "death"), method = "BFGS", control = list(maxit =
  1e+05), tcol = "times", icol = "count", debug = FALSE, ...)

Arguments

data	data frame
start	starting parameters
dist	conditional distribution of reported data
type	type of reported data
method	optimization method (see mle2)
control	control parameters for optimization
tcol	column name for time variable
icol	column name for count variable
debug	print debugging output?
...	Further arguments to pass to optimizer
optimizer	optimizer to be used (see mle2)

See Also

startfun mle2

Examples

harbin2 <- setNames(harbin,c("times","count"))
(f1 <- fitsir(harbin2, 
              start=c(beta=2, gamma=1, N=2e3, i0=0.0001, sigma=10),
              type="death"))
plot(f1)

## CRUDE R^2 analogue (don't trust it too far!)
ss <- SIR.detsim(harbin2$times,trans.pars(coef(f1)), type="death")
cc <- harbin2$count

cor(ss, cc)^2

f1_g2 <- fitsir(harbin2, 
             start=c(beta=2, gamma=1, N=2e3, i0=0.0001),
              family="gaussian2",
              type="death")
plot(f1)

bbolker/fitsir documentation built on June 4, 2019, 8:28 a.m.