control.fixed

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.

control.fixed(
  cdf = NULL,
  quantiles = NULL,
  expand.factor.strategy = "model.matrix",
  mean = 0,
  mean.intercept = 0,
  prec = 0.001,
  prec.intercept = 0,
  compute = TRUE,
  correlation.matrix = FALSE,
  remove.names = NULL
)

inla.set.control.fixed.default(...)

Arguments

cdf

A list of values to compute the CDF for, for all fixed effects

quantiles

A list of quantiles to compute for all fixed effects

expand.factor.strategy

The strategy used to expand factors into fixed effects based on their levels. The default strategy is us use the model.matrix-function for which NA's are not allowed (expand.factor.strategy="model.matrix") and levels are possible removed. The alternative option (expand.factor.strategy="inla") use an inla-specific expansion which expand a factor into one fixed effects for each level, do allow for NA's and all levels are present in the model. In this case, factors MUST BE factors in the data.frame/list and NOT added as .+factor(x1)+. in the formula only.

mean

Prior mean for all fixed effects except the intercept. Alternatively, a named list with specific means where name=default applies to unmatched names. For example control.fixed=list(mean=list(a=1, b=2, default=0)) assign 'mean=1' to fixed effect 'a' , 'mean=2' to effect 'b' and 'mean=0' to all others. (default 0.0)

mean.intercept

Prior mean for the intercept (default 0.0)

prec

Default precision for all fixed effects except the intercept. Alternatively, a named list with specific means where name=default applies to unmatched names. For example control.fixed=list(prec=list(a=1, b=2, default=0.01)) assign 'prec=1' to fixed effect 'a' , 'prec=2' to effect 'b' and 'prec=0.01' to all others. (default 0.001)

prec.intercept

Default precision the intercept (default 0.0)

compute

Compute marginals for the fixed effects ? (default TRUE)

correlation.matrix

Compute the posterior correlation matrix for all fixed effects? (default FALSE) OOPS: This option will set up appropriate linear combinations and the results are shown as the posterior correlation matrix of the linear combinations. This option will imply control.inla=list(lincomb.derived.correlation.matrix=TRUE).

remove.names

A vector of names of expanded fixed effects to remove from the model-matrix. This is an expert option, and should only be used if you know what you are doing.

...

Named arguments passed on to the main function