h.map.maker: Hurst exponent to weighted AR(1) sum mapping

In eirikmn/INLA.climate.test: Full Bayesian analysis of radiative forcing climate model

Description

Creates a mapping between the Hurst exponent of a fractional Gaussian noise and the weights and lag-one parameters of the specified AR(1) mixture approximation. The mapping functions are drawn by splines based on precomputed values included in the package.

Usage

h.map.maker(m=4, lagmax=1000, model="fgn")

Arguments

m	The number of AR(1) processes used in the approximate model.
lagmax	The number of lags of which the approximate autocorrelation functions are supposed to match the theoretical fGn one.
model	The long memory process for which H is a parameter of. Currently, only fractional Gaussian noise (model="fgn") and ARFIMA(0,d,0) (model="arfima",d=H-0.5) are supported.

Value

Returns a list of functions. The m first maps the Hurst exponent to the lag-one correlation parameters, and the last m will map to the weights.

Author(s)

Eirik Myrvoll-Nilsen [e-mail]

References

Sørbye, S., Myrvoll-Nilsen, E. and Rue, H. (2018) An approximate fractional Gaussian noise model with O(n) computational cost. Statistics and Computing.

See Also

inla.climate

Examples

m <- 4
lagmax <- 1000
H <- 0.8

h.mapping <- h.map.maker(m = m, lagmax = lagmax)

params <- numeric(m)
weights <- numeric(m)

for(i in 1:m){
  params[i] <- h.mapping[[i]](H)
  weights[i] <- h.mapping[[m+i]](H)
}

eirikmn/INLA.climate.test documentation built on Feb. 2, 2020, 9:17 a.m.