conreg-methods {cobs}R Documentation

Summary Methods for 'conreg' Objects

Description

Methods for conreg objects

Usage

## S3 method for class 'conreg'
fitted(object, ...)
## S3 method for class 'conreg'
residuals(object, ...)
## S3 method for class 'conreg'
knots(Fn, ...)

## S3 method for class 'conreg'
lines(x, type = "l", col = 2, lwd = 1.5, show.knots = TRUE,
      add.iSpline = TRUE, force.iSpl = FALSE, ...)

## S3 method for class 'conreg'
plot(x, type = "l", col = 2, lwd = 1.5, show.knots = TRUE,
    add.iSpline = TRUE, force.iSpl = FALSE,
    xlab = "x", ylab = expression(s[c](x)),
    sub = "simple concave regression", col.sub = col, ...)

## S3 method for class 'conreg'
predict(object, x, deriv = 0, ...)

## S3 method for class 'conreg'
print(x, digits = max(3, getOption("digits") - 3), ...)

Arguments

object, Fn, x

an R object of class conreg, i.e., typically the result of conreg(..). For predict(), x is a numeric vector of abscissa values at which to evaluate the concave/convex spline function.

type, col, lwd, xlab, ylab, sub, col.sub

plotting arguments as in plot.default.

show.knots

logical indicating the spline knots should be marked additionally.

add.iSpline

logical indicating if an interpolation spline should be considered for plotting. This is only used when it is itself concave/convex, unless force.iSpl is TRUE.

force.iSpl

logical indicating if an interpolating spline is drawn even when it is not convex/concave.

deriv

for predict, integer specifying the derivate to be computed; currently must be 0 or 1.

digits

number of significant digits for printing.

...

further arguments, potentially passed to methods.

Author(s)

Martin Maechler

See Also

conreg, ....

Examples

example(conreg, echo = FALSE)
class(rc) # "conreg"
rc # calls the print method
knots(rc)
plot(rc)
## and now _force_ the not-quite-concave cubic spline :
plot(rc, force.iSpl=TRUE)

xx <- seq(-0.1, 1.1, length=201) # slightly extrapolate
## Get function s(x)  and first derivative  s'(x) :
yx <- predict(rc, xx)
y1 <- predict(rc, xx, deriv = 1)

op <- par(las=1)
plot(xx, yx, type = "l",
     main="plot(xx,  predict( conreg(.), xx))")
par(new=TRUE) # draw the first derivative "on top"
plot(xx, y1, type = "l", col = "blue",
     axes = FALSE, ann = FALSE)
abline(h = 0, lty="1A", col="blue")
axis(4, col="blue", col.axis="blue", col.ticks="blue")
mtext("first derivative s'(.)", col="blue")
par(op)

[Package cobs version 1.3-8 Index]