| dlmBSample {dlm} | R Documentation |
The function simulates one draw from the posterior distribution of the state vectors.
dlmBSample(modFilt)
modFilt |
A list, typically the ouptut from dlmFilter, with elements
m, U.C, D.C, a, U.R, D.R (see
the value returned by dlmFilter), and mod The latter is an
object of class "dlm" or a list with elements GG, W
and, optionally, JGG, JW, and X |
The calculations are based on singular value decomposition.
The function returns a draw from the posterior distribution
of the state vectors. If m is a time series then the returned
value is a time series with the same tsp, otherwise it is
a matrix or vector.
Giovanni Petris GPetris@uark.edu
West and Harrison, Bayesian forecasting and dynamic models (2nd ed.), Springer (1997).
See also dlmFilter
nileMod <- dlmModPoly(1, dV = 15099.8, dW = 1468.4)
nileFilt <- dlmFilter(Nile, nileMod)
nileSmooth <- dlmSmooth(nileFilt) # estimated "true" level
plot(cbind(Nile, nileSmooth$s[-1]), plot.type = "s",
col = c("black", "red"), ylab = "Level",
main = "Nile river", lwd = c(2, 2))
for (i in 1:10) # 10 simulated "true" levels
lines(dlmBSample(nileFilt[-1]), lty=2)