Plot (restricted) ROC curves — plot_rROC

Plot for every tpr/fpr combination:

Full density
Full ROC + selection rectangles (orange + blue) + Full AUC + orange+blue partial AUCs + scaling factor + name rectangle sides + orange+blue partial AUCs SCALED
ROCs for selections + Full AUCs + Color rectangle around the plot
Densities for selections

Usage

plot_rROC_part(
  x,
  p_full_density_ROC = NA,
  threshold = NA,
  fpr = NA,
  color_high = default_part_colors["high"],
  color_low = default_part_colors["low"],
  include_part_auc_text = FALSE
)

Arguments

x: rROC result including full_pROC.
p_full_density_ROC: Density plot from plot_density_ROC_empirical of the full data, if not given it is calculated from rROC_res
threshold: At which threshold should be split into high and low. Alternatively, give a false positive rate (fpr)
fpr: A false positive rate at which (approximately) the cutoff will be made
color_high: Color for high part
color_low: Color for low part
include_part_auc_text: Should the text-annotations be added or not?

Value

patchworked plots, see the description.

Examples


library(restrictedROC)
data(aSAH, package = "pROC")
ret_procs <- simple_rROC(
    response = aSAH$outcome,
    predictor = aSAH$ndka,
    return_proc = TRUE
)
#> Positive label not given, setting to last level of factor: Poor
# pdf("removeme.pdf")
print(plot_rROC_part(ret_procs, fpr = .5))
#> $plotlist
#> $plotlist$A

#> 
#> $plotlist$B

#> 
#> $plotlist$C

#> 
#> $plotlist$D

#> 
#> $plotlist$E

#> 
#> $plotlist$F

#> 
#> 
#> $patchworked

#> 
# dev.off()