covRob
.
covRob.control(estim, quan = NULL, ntrial = 500, r = 0.45, alpha = 0.05, tau = 1e-06, tol = 0.001, maxit = 150, nresamp = "Auto", maxres = "Auto", random.sample = F, tune = 0.95, prob = 0.99, eps = 0.5)
the robust estimator used in covRob. The choices are: "mcd" for the Fast
MCD algorithm of Rousseeuw and Van Driessen, "donostah" for the
Donoho-Stahel projection based estimator, "M" for the constrained M
estimator provided by Rocke, "pairwiseQC" for the quadrant correlation
based pairwise estimator, and "pairwiseGK" for the Gnanadesikan-Kettenring
pairwise estimator. Only the parameters used by
estim
are returned.
quan
is
floor((n+p+1)/2)
, where
n
is the number of observations and
p
is the number of variables. Any
quan
between the default and
n
may be specified. If
quan
is between 0.5 and 1.0 then it is
interpreted as the fraction of observations that should be used.
the fraction of contaminated data. The default value is 0.45.
the number of resamples required;
nresamp
may
not be reached if too many of the size
'p'
subsamples, chosen out of the observed vectors, are in a hyperplane. If
nresamp=0
, ALL subsamples are taken. The
default
nresamp
is calculated to ensure a
breakdown point of
eps
with probability
prob
.
"Auto"
is 2 times
nresamp
.
prob
is the probability of high breakdown. The
default value is .99.
eps
is the fraction of contamination in the
data. The default value is .5.
covRob.control.list <- covRob.control(estim = "M", r = .2)
covRob(woodmod.dat, estim = "M", control = covRob.control.list)