In general, some assembly is required. This stuff doesn't quit fit in a one
page email. You may want to start playing around with model.matrix and qr() to
see the light.
One tricky bit is how much you know already. There is some risk for me of
having to rewrite a linear algebra textbook...
I'l
Hi Peter,
Thank you for your good answer. I am sorry for the late reply.
*An ortogonalized model matrix generates a decomposition of the model space
into orthogonal subspaces corresponding to the terms of the model.
Projections onto each of the subspaces are easily worked out. E.g., for a
two-w
I am not aware of a detailed documentation of this beyond the actual source
code.
However, the principles are fairly straightforward, except that the rules for
constructing the design matrix from the model formula can be a bit arcane at
times.
The two main tools are the design matrix construct
Hi
*I have been looking for a reference to explain how R uses the aov
command(at a deeper level)*. More specifically, how R reads the formulae
and R computes the sums of squares. I am not interested in understanding
what the difference of Type 1,2,3 sum of squares are. I am more interested
in find
4 matches
Mail list logo
