plot.isoreg {stats}R Documentation

Plot Method for isoreg Objects


The plot and lines method for R objects of class isoreg.


## S3 method for class 'isoreg'
plot(x, plot.type = c("single", "row.wise", "col.wise"),
      main = paste("Isotonic regression", deparse(x$call)),
      main2 = "Cumulative Data and Convex Minorant",
      xlab = "x0", ylab = "x$y", = list(col = "red", cex = 1.5, pch = 13, lwd = 1.5),
      mar = if (both) 0.1 + c(3.5, 2.5, 1, 1) else par("mar"),
      mgp = if (both) c(1.6, 0.7, 0) else par("mgp"),
      grid = length(x$x) < 12, ...)

## S3 method for class 'isoreg'
lines(x, col = "red", lwd = 1.5,
       do.points = FALSE, cex = 1.5, pch = 13, ...)



an isoreg object.


character indicating which type of plot is desired. The first (default) only draws the data and the fit, where the others add a plot of the cumulative data and fit. Can be abbreviated.


main title of plot, see title.


title for second (cumulative) plot.

xlab, ylab

x- and y- axis annotation.

a list of arguments (for points and lines) for drawing the fit.

mar, mgp

graphical parameters, see par, mainly for the case of two plots.


logical indicating if grid lines should be drawn. If true, grid() is used for the first plot, where as vertical lines are drawn at ‘touching’ points for the cumulative plot.


for lines(): logical indicating if the step points should be drawn as well (and as they are drawn in plot()).

col, lwd, cex, pch

graphical arguments for lines(), where cex and pch are only used when do.points is TRUE.


further arguments passed to and from methods.

See Also

isoreg for computation of isoreg objects.



utils::example(isoreg) # for the examples there

plot(y3, main = "simple plot(.)  +  lines(<isoreg>)")

## 'same' plot as above, "proving" that only ranks of 'x' are important
plot(isoreg(2^(1:9), c(1,0,4,3,3,5,4,2,0)), plot.type = "row", log = "x")

plot(ir3, plot.type = "row", ylab = "y3")
plot(isoreg(y3 - 4), plot.type = "r", ylab = "y3 - 4")
plot(ir4, plot.type = "ro",  ylab = "y4", xlab = "x = 1:n")

## experiment a bit with these (C-c C-j):
plot(isoreg(sample(9),  y3), plot.type = "row")
plot(isoreg(sample(9),  y3), plot.type = "col.wise")

plot(ir <- isoreg(sample(10), sample(10, replace = TRUE)),
                  plot.type = "r")

[Package stats version 4.4.0 Index]