sim_sfda_curves: simulates a collection of curves with spatial dependendence.

Description Usage Arguments Value Examples

Description

For the specified basis functions, spatial dependence is incorporated through the covariance structure of the spatial fields connected to the coefficients. The coefficients of the first basis funciton is a simulated Gaussian random field with covariance model equal to the first specified covariance model and parameters equal to the first specified covariance parameters. The same is true for the remaining basis functions, which means that the number of basis functions should be equal to the number of covariance models specified, and the order in which they are specified is matters. The Gaussian random fields are simulated with the grf function in the geoR package. See geoR documentation for specification of covariance models.

Usage

1
2
3
sim_sfda_curves(nBasis, cov.model, cov.pars, ..., type = "Cos",
  basis.pars = NULL, grid.dim = c(10, 10), grid.xlim = c(0, 1),
  grid.ylim = c(0, 1), locs = NULL, write = FALSE, file = NULL)

Arguments

nBasis

number of basis functions

cov.model

character vector specifying covariance function(s). The length of this vector should equal the number of basis functions.

cov.pars

parameters for the covariance function(s) specified in cov.model

...

additional arguments sent to grf. See help for grf in goeR for details.

type

character specifying the type of basis functions to use. Default is type="Cos".

basis.pars

extra parameters used in the construction of basis functions

grid.dim

vector with 2 numbers specifying the dimension of the rectangular grid. Only used if locs is not specified.

grid.xlim

vector of length 2 specifying the xlim for the grid. Only used if locs is not specified.

grid.ylim

vector of length 2 specifying the ylim for the grid. Only used if locs is not specified.

locs

nx2 matrix of locations. If NULL, then grid values are used to create locations

write

logical. If write=TRUE will write results to file

file

character specifying the file name. Only used if write = TRUE

Value

locs matrix containing spatial locations of the curves

coef matrix containing coefficients of the basis functions. The number of columns is equal to the number of basis functions

basis.fns list of basis functions created by create_basis

Examples

1
2
3
4
5
curves <- sim_sfda_curves(nBasis = 2, cov.model = c("gaussian", "exponential"), cov.pars = rbind(c(1, 0.5), c(1, .3)), type="Cos", basis.pars = 2, locs = expand.grid(1:5/5, 1:10/10))
plot_curves(curves$coef, curves$basis.fns, ylim=c(-3,3))

sim.data <- sim_sfda_data(curves$locs, curves$coef, curves$basis.fns, sigma0=0.4, m = 10)
## plot_data(sim.data)

dan410/sfdasim documentation built on May 14, 2019, 3:34 p.m.