Prediction with splines

From: Kevin Wright <kw.statr_at_GMAIL.COM>
Date: Tue, 6 Nov 2007 16:32:49 -0600

David, I have a question about a quirk with predictions in asreml-r.
I thought about asking this on the mailing list, but other people may
not be interested in this particular question. If you prefer that I
ask this on the mailing list so that it is archived, just let me know.

Here is the situation:

tree1 <- subset(orange, Tree==1)
m1 <- asreml(circ ~ 1 + lin(x) , random= ~ spl(x), data=tree1)
m1 <- update(m1)
p1 <- predict(m1, classify="x",
              predictpoints=list(x=seq(from=118, to=1582, length=20)))
p1 <- p1$predictions$x$pvals

- The cells of the hypertable are calculated from all model terms constructed
  solely from factors in the averaging and classify sets.

           x predicted.value standard.error est.status
1 118.0000 30.76666 6.044057 Estimable
2 195.0526 37.89921 5.270393 Estimable
3 272.1053 45.02535 4.689576 Estimable
4 349.1579 52.13864 4.312880 Estimable

My question is: sometimes in the "Notes" section of the predicted
values I see output similar to:
"lin(x) evaluated at average value of 5.500000"

This time, however, there is no such note. I have not figured out
when I see this note about 'x evaluated at average value' and when I

Is there a reason for this that I have overlooked, or maybe this is
just missing in some (which?) models?


Received on Sat Nov 06 2007 - 16:32:49 EST

This webpage is part of the ASReml-l discussion list archives 2004-2010. More information on ASReml can be found at the VSN website. This discussion list is now deprecated - please use the VSN forum for discussion on ASReml. (These online archives were generated using the hypermail package.)