remotePARTS (version 1.0.4)

optimize_nugget: Find the maximum likelihood estimate of the nugget

Description

Find the maximum likelihood estimate of the nugget

Usage

optimize_nugget(
  X,
  y,
  V,
  lower = 0.001,
  upper = 0.999,
  tol = .Machine$double.eps^0.25,
  debug = FALSE,
  ncores = NA
)

Value

maximum likelihood nugget estimate

Arguments

X

numeric (double) nxp matrix

y

numeric (double) nx1 column vector

V

numeric (double) nxn matrix

lower

lower boundary for nugget search

upper

upper boundary for nugget search

tol

desired accuracy of nugget search

debug

logical: debug mode?

ncores

an optional integer indicating how many CPU threads to use for matrix calculations.

Details

Finds the maximum likelihood nugget estimate via mathematical optimization.

To maximize efficiency, optimize_nugget() is implemented entirely in C++. Optimization takes place via a C++ version of the fmin routine (Forsythe et al 1977). Translated from http://www.netlib.org/fmm/fmin.f

The function LogLikGLS() is optimized for nugget. Once the LogLikGLS() functionality is absorbed by fitGLS(), it will be used instead.

See Also

?stats::optimize()