control.expert {INLA}R Documentation

control.expert

Description

Control variables in ⁠control.*⁠ for use with inla(). The functions can be used to TAB-complete arguments, and returns a list of the default control arguments, unless overridden by specific input arguments.

Usage

control.expert(
  cpo.manual = FALSE,
  cpo.idx = -1,
  disable.gaussian.check = FALSE,
  jp = NULL,
  dot.product.gain = FALSE,
  globalconstr = list(A = NULL, e = NULL),
  opt.solve = FALSE,
  opt.num.threads = TRUE
)

inla.set.control.expert.default(...)

Arguments

cpo.manual

A boolean variable to decide if the inla-program is to be runned in a manual-cpo-mode. (EXPERT OPTION: DO NOT USE)

cpo.idx

The index/indices of the data point(s) to remove. (EXPERIMENTAL OPTION: DO NOT USE)

disable.gaussian.check

Disable the check for fast computations with a Gaussian likelihood and identity link (default FALSE)

jp

An object of class inla.jp defining a joint prior

dot.product.gain

Output the gain in optimizing dot-products? (Default FALSE). This options needs developer version, otherwise its disabled.

globalconstr

Add a global constraint (see ?f and argument extraconstr). Note that a global constraint does NOT correct the normalisation constant. (EXPERIMENTAL OPTION)

opt.solve

Store also L^T to optimize linear solves (TAUCS only). (EXPERIMENTAL OPTION: DO NOT USE)

opt.num.threads

Reduce the number of threads adaptively if needed at critical places to optimize wall-clock time (EXPERIMENTAL OPTION)

...

Named arguments passed on to the main function

See Also

Other control: control.bgev(), control.compute(), control.family(), control.fixed(), control.gcpo(), control.group(), control.hazard(), control.inla(), control.lincomb(), control.link(), control.lp.scale(), control.mix(), control.mode(), control.numa(), control.pardiso(), control.pom(), control.predictor(), control.scopy(), control.sem(), control.stiles(), control.taucs(), control.update(), control.vb()


[Package INLA version 25.10.19 Index]