Plots overview

Every precrec object knows how to draw itself two ways.

library(precrec)
library(ggplot2)

curves <- evalmod(scores = P10N10$scores, labels = P10N10$labels)

plot or autoplot

Function	Draws with	Use it when
plot()	base graphics	You want a picture with no dependencies
autoplot()	ggplot2	You want to restyle, relabel or combine the result

plot(curves)

autoplot() returns a ggplot object, so anything you can do to a plot from ggplot2 you can do to this one.

autoplot(curves)

Choosing what to draw

The second argument picks the curve or the measure.

autoplot(curves, "PRC")

For mode = "basic" objects it names measures instead, one panel each. See basic measure plots.

Arguments both functions share

Argument	Effect
type	"l" lines, "p" points, "b" both
show_cb	Draw the confidence band around an average curve
raw_curves	Draw every test set instead of the average
show_legend	Show or hide the legend
reduce_points	Thin the points before drawing; on by default

autoplot(curves, "PRC", type = "b")

The four pages

• ROC and precision-recall curves - the main plot
• Basic measure plots - measures against normalized rank
• One measure against another - free-form pairs
• Confidence bands and partial curves - the two variations

And customizing plots for changing what comes out.