a fitted model, inheriting from lm, including those with multiple response.
OPTIONAL ARGUMENTS:
complete:
partial:
flags indicating whether information for complete and partial aliasing
should be included in the result. Defaults are TRUE
pattern:
should the resulting alias matrices be simplified by calling the pattern()
function. Default is TRUE.
VALUE:
a list potentially containing components for complete aliasing
and for partial aliasing;
each is included only if both requested and found to exist in the
model.
The component for complete aliasing is a matrix with
columns corresponding to effects that are linearly dependent on the rows
(i.e., effects that are completely aliased with the estimable effects).
Partial aliasing is essentially the correlation of the estimable effects,
with the diagonal elements set to zero.
DETAILS:
This function is a method for the generic function
alias
for class
"lm".
It can be invoked by calling
alias
for an object of the appropriate class, or directly by calling
alias.lm
regardless of the class of the object.
The method for mlm objects uses that for lm objects, after
adjusting the object suitably.