Approximate variance-compensating basis functions

Calculates an approximate basis for tau and kappa for an inla.spde2.matern model where tau is a rescaling parameter.

inla.spde2.matern.sd.basis(
  mesh,
  B.sd,
  B.range,
  method = 1,
  local.offset.compensation = FALSE,
  alpha = 2,
  ...
)

Arguments

mesh

An fmesher::fm_mesh_1d(), fmesher::fm_mesh_2d(), or fmesher::fm_mesh_3d() object.

B.sd

Desired basis for log-standard deviations.

B.range

Desired basis for spatial range.

method

Construction method selector. Expert option only.

local.offset.compensation

If FALSE, only compensate in the average for the tau offset.

alpha

The model alpha parameter.

...

Additional parameters passed on to internal inla.spde2.matern calls.

Value

List of basis specifications

B.tau

Basis for log(tau)

B.kappa

Basis for log(kappa)

Intended for passing on to inla.spde2.matern().

Author

Finn Lindgren finn.lindgren@gmail.com