gremlinControl function

Advanced Options for Mixed-effect modeling functions.

Change default settings for gremlin models.

gremlinControl(
  cctol = c(5e-04, 1e-05, 0.01, NULL),
  ezero = 1e-08,
  einf = 1e+30,
  step = 0.3,
  h = .Machine$double.eps^(1/3),
  lambda = FALSE,
  algorithm = NULL,
  algArgs = list()
)

Arguments

• cctol: Convergence criteria tolerances (Meyer 2007, 2019). See details about the convergence checks in ccFun.
• ezero: Effective zero to be used, values less than this number are treated as zero and fixed to this value.
• einf: Effective infinite value to be used, values are limited to a to this variable as a maximum.
• step: A numeric value for scaling the proposed parameter updates.
• h: A numeric value for finite difference algorithm of partial first derivatives for the likelihood function. Specifies the amount to add to parameter value to calculate the change in log-likelihood.
• lambda: A logical indicating whether a residual variance should be factored out of the mixed model equations.
• algorithm: A character naming the function to use to decide subsequent parameters in the REML iterations.
• algArgs: A list of function arguments to be given to functions named in the algorithm argument.

Returns

A list of class gremlinControl to be used by gremlinSetup and later functions when fitting the model.

Examples

str(gremlinControl())

References

Meyer, K. 2007. WOMBAT - a tool for mixed model analyses in quantitative genetics by restricted maximum likelihood (REML). Journal of Zhejiang University SCIENCE B 8(11):815-821.

Meyer, K. 2019. WOMBAT A program for mixed model analyses by restricted maximum likelihood. User Notes. 27 September 2019.

Author(s)

matthewwolak@gmail.com