asympvar: Asymptotic Variance

In geoRglm: A Package for Generalised Linear Spatial Models

Description

Calculates the initial monotone/positive sequence estimate of the asymptotic variance from CLT (Geyer 92). Useful for estimation of the variance of a Markov Chain Monte Carlo estimate.

Usage

asympvar(timeseries, type="mon", lag.max = 100, messages)

Arguments

timeseries	a vector with a timeseries, or a matrix where the rows are different timeseries.
type	"pos" and "mon" gives the monotone and the positive sequence estimator, respectively, and "all" gives both. Default is type="mon".
lag.max	maximum lag at which to calculate the asymptotic variance. Default is lag.max = 100.
messages	logical. If TRUE, the default, status messages are printed while the function is running.

Value

A number (or a vector) with the estimate, when type="mon" or type="pos". A list with components mon and pos when type="all"

Author(s)

The orginal Splus version of this function was written by Rasmus Waagepetersen. R port by
Ole F. Christensen [e-mail],
Paulo J. Ribeiro Jr. [e-mail].

References

Geyer, C. (1992). Practical Monte Carlo (with discussion). Statist. Sci. 7, 473-511.

Further information about geoRglm can be found at:
http://gbi.agrsci.dk/~ofch/geoRglm.

Examples

data(p50)
## Not run: 
test <- pois.krige(p50, krige = krige.glm.control(cov.pars = c(1,1), beta = 1),
      mcmc.input = mcmc.control(S.scale = 0.5, n.iter = 1000, thin = 1))
asympvar(test$intensity[45,])
ass <- asympvar(test$intensity[1:10,], type = "pos")

## End(Not run)

geoRglm documentation built on May 2, 2019, 4:03 p.m.