lts
They represent a robust fit of a linear model.
This class of objects is returned from the
ltsreg function.
The
"lts" class of objects has methods for the following generic
functions:
plot,
print,
summary.
The following components must be included in a legitimate
lts object.
singular.ok is
TRUE in the call to
ltsreg and
x is singular,
then some of the elements will be
NA.
floor((n+p+1)/2), where
n is the
number of observations and
p the number of variables.
quan smallest squared residuals) over
(the sum of the
quan smallest
(y-loc)^2),
where the denominator is minimized over
loc.
Note that
loc is not subtracted from
y if
intercept = FALSE in the
call to
ltsreg.
y containing the residuals from the regression.
Rows corresponding to missing values in
x or
y have
NAs.
y containing weights that can be used in weighted least squares.
These weights are
1 for points with reasonably small residuals,
0
for points with large residuals.
This is only included if
wt is
TRUE in the call to
ltsreg.
intercept=TRUE).
mcd=TRUE),
the robust distances of the
x-cases obtained from
cov.mcd
are returned.
x in a call to
ltsreg.default,
or the data matrix.
y containing the fitted values.
nkeep containing vectors of observation numbers that define
fits.
qr on
x.
This is only present when
qr.out is
TRUE in the call to
ltsreg.
model=TRUE (in
ltsreg.formula).
x=TRUE (in
ltsreg.formula).
y=TRUE (in
ltsreg.formula).