Dear list,
When I use the predict.glm method on a glm fitted object, I get the
following warning message:
In addition: Warning message:> In predict.lm(object, x) :
> prediction from a rank-deficient fit may be misleading
As the documentation says this happens if "the fit is rank-deficient, some
of the columns of the design matrix will have been dropped".
It is my understanding that this means that at least one predictor variable
in the design matrix can be replicated by a linear combination of the other
predictors. But this doesn't seem the case in my data set. I noticed that
the problem goes away after removing higher order interaction terms from the
model. Some of the coefficients from those terms were NA as shown in the
summary(glm).
I'm wondering if my interpretation of the message is correct, and/or in what
type of situations would I get this warning message.
Thanks in advance for your help.
Lars.
[[alternative HTML version deleted]]
