estcovparm function

Estimate Covariance Parameters

Used to estimate spatial covariance parameters for a few different spatial models. Estimated parameters can then be used in predict.slmfit() to predict values at unobserved locations.

estcovparm(
  response,
  designmatrix,
  xcoordsvec,
  ycoordsvec,
  CorModel = "Exponential",
  estmethod = "REML",
  covestimates = c(NA, NA, NA)
)

Arguments

• response: a vector of a response variable, possibly with missing values.
• designmatrix: is the matrix of covariates used to regress the response on.
• xcoordsvec: is a vector of x coordinates
• ycoordsvec: is a vector of y coordinates
• CorModel: is the covariance structure. By default, CorModel is "Exponential" but other options are "Spherical" and "Gaussian".
• estmethod: is either the default "REML" for restricted maximum likelihood to estimate the covariance parameters and regression coefficients or "ML" to estimate the covariance parameters and regression coefficients.
• covestimates: is an optional vector of covariance parameter estimates (nugget, partial sill, range). If these are given and estmethod = "None", the the provided vector are treated as the estimators to create the covariance structure.

Returns

a list with

• parms.est, a vector of estimated covariance parameters
• Sigma, the fitted covariance matrix for all of the sites
• qrV, the qr decomposition
• b.hat, the vector of estimated fixed effect coefficients
• covbi, the inverse of the covariance matrix for the fixed effects
• covb, the covariance matrix for the fixed effects
• min2loglik, minus two times the loglikelihood

Details

The function is a helper function used internally in predict.slmfit().