Spline predict points

From: Kevin Wright <kw.statr_at_GMAIL.COM>
Date: Fri, 26 Oct 2007 13:39:06 -0500

I'm exploring the use of splines with a mixed model in asreml-r
following the example of the orange tree growth for Tree 1.

Page 117 of ASREML-R manual says:
"The predictpoints argument adds the nominated points to the design
matrix for prediction purposes (Figure 8.13)."

Does this mean that I will obtain predictions at the design points AND
the predictpoints, or only at predictpoints, or something else?

Here's a simple example:

o1 <- asreml(circ ~ 1 + lin(x),
random= ~spl(x, 7),
splinepoints=list(x=c(118, 484, 664, 1004, 1231, 1372, 1582)),
data=orange, subset=Tree==1)
o.pred <- predict(o1, classify = "x", predictpoints=list(x=200))
o.pred\$predictions\$x\$pvals

Notes:
- 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
1 118.0 30.76677 6.044060 Estimable
2 191.2 37.54277 5.304859 Estimable
3 264.4 44.31327 4.738176 Estimable
4 337.6 51.07276 4.357754 Estimable
5 410.8 57.81574 4.126452 Estimable
6 484.0 64.53670 3.978459 Estimable
7 557.2 71.22233 3.880751 Estimable
8 630.4 77.82801 3.849271 Estimable
9 664.0 80.81897 3.846828 Estimable
10 737.2 87.22137 3.829022 Estimable
etc

Since I specified a prediction at x=200, I was expecting to see one
row for "x=200" in the table.

Would appreciate any clarifications of this.

Thanks, K Wright
Received on Tue Oct 26 2007 - 13:39:06 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.)