OsplinePen : O-spline penalty matrix

Description

This function returns the penalty matrix for smoothing spline of any order.

Usage

OsplinePen(Boundary.knots, knots, ord=1)

Arguments

Boundary.knots

A length 2 numeric vector, giving the boundary knot values.

knots

A nuemric vector of internal knots.

ord

A numeric integer, which is the order of the derivatives on which squared integral will become the smoothness penalty.

Value

A symmetrix penalty matrix.

References

Wand, M. P. and Ormerod, J. T. (2008) On semiparametric regression with O'Sullivan penalized splines. Aust. N. Z. J. Stat. 50(2), 179--198.

Examples

Run this code

# NOT RUN {
    b.k=c(0,1)
    br=seq(.1,.9,by=.1)
    O1=OsplinePen(b.k, br, 1)
    O2=OsplinePen(b.k, br, 2)
    O3=OsplinePen(b.k, br, 3)
    O4=OsplinePen(b.k, br, 4)
    O5=OsplinePen(b.k, br, 5)
    O6=OsplinePen(b.k, br, 6)

    library(fda)
# }
# NOT RUN {
    des1=create.bspline.basis(c(0,1),norder=2, breaks=br)
    P1=bsplinepen(des1, 1) # ERROR
    max(abs(P1-O1))
# }
# NOT RUN {
    des2=create.bspline.basis(c(0,1),norder=4, breaks=c(b.k[1], br, b.k[2]))
    P2=bsplinepen(des2, 2)
    max(abs(P2-O2))

    des3=create.bspline.basis(c(0,1),norder=6, breaks=c(b.k[1], br, b.k[2]))
    P3=bsplinepen(des3, 3)
    max(abs(P3-O3))

    des4=create.bspline.basis(c(0,1),norder=8, breaks=c(b.k[1], br, b.k[2]))
    P4=bsplinepen(des4, 4, c(0,1))
    max(abs((P4-O4)/(P4+O4)*2),na.rm=TRUE)

    des5=create.bspline.basis(c(0,1),norder=10, breaks=c(b.k[1], br, b.k[2]))
    P5=bsplinepen(des5, 5, c(0,1))
    max(abs((P5-O5)/(P5+O5)*2),na.rm=TRUE)

    des6=create.bspline.basis(c(0,1),norder=12, breaks=c(b.k[1], br, b.k[2]))
    P6=bsplinepen(des6, 6, c(0,1))
    max(abs((P6-O6)/(P6+O6)*2),na.rm=TRUE)

# }