inla.spde1.create: Old SPDE model objects for INLA

In andrewzm/INLA: Functions which allow to perform full Bayesian analysis of latent Gaussian models using Integrated Nested Laplace Approximaxion

Description

Create an inla.spde1 model object.

Usage

## Create an SPDE model object:
inla.spde1.create(mesh,
                  model = c("matern", "imatern", "matern.osc"),
                  param = NULL,
                  ...)

## Shortcuts to the matern, imatern and matern.osc models:
inla.spde1.matern(mesh, ...)
inla.spde1.imatern(mesh, ...)
inla.spde1.matern.osc(mesh, ...)

Arguments

mesh	The mesh to build the model on, as an inla.mesh object.
model	The name of the model.
param	Model specific parameters.
...	Additional parameters passed on to other methods.

Details

Note: This is an old spde object format retained for backwards compatibility. Please use inla.spde2 models for new code.

This method constructs an object for SPDE models. Currently implemented:

model="matern"

(kappa^2(u)-Delta)^(alpha/2) (tau(u) x(u)) = W(u)

param:

• alpha = 1 or 2

• basis.T = Matrix of basis functions for log tau(u)

• basis.K = Matrix of basis functions for log kappa^2(u)

model="imatern"

(-Delta)^(alpha/2) (tau(u) x(u)) = W(u)

param:

• alpha = 1 or 2

• basis.T = Matrix of basis functions for log tau(u)

Value

An inla.spde1 object.

Author(s)

Finn Lindgren [e-mail]

See Also

inla.spde2.matern, inla.mesh.2d, inla.mesh.basis

Examples

n = 100
field.fcn = function(loc) (10*cos(2*pi*2*(loc[,1]+loc[,2])))
loc = matrix(runif(n*2),n,2)
## One field, 2 observations per location
idx.y = rep(1:n,2)
y = field.fcn(loc[idx.y,]) + rnorm(length(idx.y))

mesh = inla.mesh.create(loc, refine=list(max.edge=0.05))
spde = inla.spde.create(mesh, model="matern")
data = list(y=y, field=mesh$idx$loc[idx.y])
formula = y ~ -1 + f(field, model=spde)
result = inla(formula, data=data, family="normal")

## Plot the mesh structure:
plot(mesh)

if (require(rgl)) {
  ## Plot the posterior mean:
  plot(mesh, rgl=TRUE,
       result$summary.random$field[,"mean"],
       color.palette = colorRampPalette(c("blue","green","red")))
  ## Plot residual field:
  plot(mesh, rgl=TRUE,
       result$summary.random$field[,"mean"]-field.fcn(mesh$loc),
       color.palette = colorRampPalette(c("blue","green","red")))
}

andrewzm/INLA documentation built on May 10, 2019, 11:12 a.m.