predict.smooth.Pspline {pspline} | R Documentation |
Smoothing Spline of Arbitrary Order at New Data
Description
Uses an object of class "smooth.Pspline"
to evaluate a
polynomial smoothing spline of arbitrary order or one of its
derivatives at new argument values.
Usage
## S3 method for class 'smooth.Pspline'
predict(object, xarg, nderiv = 0, ...)
Arguments
object |
a fitted |
xarg |
the argument values at which the spline or its derivative is to be evaluated. |
nderiv |
the order of the derivative required – the default is 0, the function itself. |
... |
further arguments passed to or from other methods. |
Details
The method produces results similar to function the predict
method for smooth.spline
, but the smoothing function is
a natural smoothing spline rather than a B-spline smooth, and the
order of the spline can be chosen freely, where order in this case
means the order of the derivative that is
penalized. smooth.spline
penalizes the second derivative, and
consequently only derivatives or order 0 or 1 are useful, but because
smooth.Pspline
penalizes a derivative of order m,
derivatives up to order m-1 are useful. The general recommendation is
to penalize the derivative two beyond the highest order derivative to
be evaluated.
Value
A list with components xarg
and dy
; the xarg
component is identical to the input xarg
sequence, the
dy
component is the evaluated derivative of order deriv
.
References
Heckman, N. and Ramsay, J. O. (1996) Spline smoothing with model based penalties. McGill University, unpublished manuscript.
See Also
Examples
example(smooth.Pspline)
## smoother line is given by
xx <- seq(4, 25, length=100)
lines(xx, predict(sm.spline(speed, dist, df=5), xx), col = "red")
## add plots of derivatives
lines(xx, 10*predict(sm.spline(speed, dist), xx, 1), col = "blue")
lines(xx, 100*predict(sm.spline(speed, dist), xx, 2), col = "green")