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
NA
s.
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
).