balanced: Compute the M-quantile extension of the mixed linear model...

In mqext: M-quantile extension of the mixed linear model

Description

Compute the M-quantile extension of the mixed linear model for BALANCED DATA ONLY

Usage

balanced(model, k, q, init = list(d = 1, v = 1), store = FALSE, 
         acc = 1e-05, niter = 30, nonzero = TRUE)

Arguments

model	a saemodel object
k	scalar; robustness-tuning constant of the asymmetric Huber psi-function
q	scalar; quantile (0,1);
init	initial values
store	logical; if store=TRUE, the function returns the iteration-specific estimates
acc	scalar; accuarcy measure used in the termination rule of the algorithm
niter	scalar; number of iterations
nonzero	logical; if nonzero=FALSE, the method also uses negative variance estimates (sic!)

Details

This function computes M-quantile-type estimates of the fixed effects and the random-effect variances of the balanced-data mixed linear model.

The model must be defined by means of the saemodel function in the rsae package.

Value

beta	vector fixed-effects estimates
s_e	model-error std. dev. (not variance)
s_a	area-level-random-effect std. dev. (not variance)
niter	number of iterations used to compute the estimates

Author(s)

Tobias Schoch

References

Tzavidis et al. (2011)

mqext documentation built on May 2, 2019, 4:52 p.m.