fitsaemodel.control: Tuning parameters of fitsaemodel
In rsae: Robust Small Area Estimation

Description Usage Arguments Details Value Author(s) See Also

This function carries global settings and parameter definitions that are used by fitsaemodel (and its derivatives). Modifications of the parameters can be delivered as additional arguments in the fitsaemodel call.

1 2	fitsaemodel.control(niter = 40, iter = c(200, 200), acc = 1e-05, dec=0, decorr=0, init="default", ...)

`niter`	integer, defining the maximum number of outer-loop iterations (default: `niter=40`)
`iter`	integer or vector of size 2, defining the maximum loops of the inner loops (default: `iter=c(200, 200)`; element 1 refers to beta; element 2 refers to v; note that d has an implicitly defined maxiter of 100 and cannot be modified)
`acc`	scalar or vector of size 4, defining the numeric tolerance used in the termination rule of the loops (default: `acc=1e-05`; the positions of elements in the vector of size 4 are: 1=acc outer-loop; 2=acc inner-loop beta; 3=acc inner-loop v; 4=acc inner-loop d).
`dec`	type of matrix square root (decomposition); `dec=0` for SVD (default) or `dec=1` for Cholesky decomposition
`decorr`	in order to robustly decorrelate the residuals, one has to chose `decorr=1` (default `decorr=0`)
`init`	a character string; specifies by what method the main algorithm is initialized; by default: `init="default"`; alternatively, (and provided that the robustbase package is installed) one may choose a high-breakdown-point initial estimate: either "lts" (fast LTS regression) or "s" (S-estimate of regression). For datasets with more than 100,000 observations, the former is rather slow. For more details on the initializing methods see the documentation of robustbase (`"ltsReg"` and `"lmrob.S"`).
`...`	(will be used in the future)