seffEst {stfit}

STFIT Spatial Effect Estimation

Description

STFIT Spatial Effect Estimation

Usage

seffEst(
  rmat,
  img.nrow,
  img.ncol,
  h.cov = 2,
  h.sigma2 = 2,
  weight.cov = NULL,
  weight.sigma2 = NULL,
  nnr,
  method = c("lc", "emp"),
  partial.only = TRUE,
  pve = 0.99,
  msk = NULL,
  msk.tol = 0.95,
  var.est = FALSE
)

Arguments

rmat

residual matrix

img.nrow

image row dimension

img.ncol

image column dimension

h.cov

bandwidth for spatial covariance estimation; ignored if weight.cov is supplied

h.sigma2

bandwidth for sigma2 estimation

weight.cov

weight matrix for spatial covariance estimation

weight.sigma2

weight vector for spatial variance estimation

nnr

maximum number of nearest neighbor pixels to use for spatial covariance estimation

method

"lc" for local constant covariance estimation and "emp" for empirical covariance estimation

partial.only

calculate the spatical effect for partially observed images only, default is TRUE

pve

percent of variance explained of the selected eigen values. Default is 0.99.

msk

an optional logistic vector. TRUE represent the corresponding pixel is always missing.

msk.tol

if 'msk' is not given, the program will determine the mask using getMask function. If the percentage of missing values for a pixel over time is greater than this

var.est

Whether to estimate the variance of the temporal effect. Default is FALSE.

Value

List of length 3 with entries:

• seff_mat: estimated spatial effect matrix of the same shape as rmat.

• seff_var_mat: estimated spatial effect variance matrix of the same shape as rmat.

• idx: a list of two entries:

  • idx.allmissing: index of the completely missing images.

  • idx.imputed: index of the partially observed images, where spatial effects are estimated.