gg_CV: Plot confusion matrices

In MomX/Momstats: Statistical Methods on top of Momocs

Description

Plots as heatmap and with useful cosmetics

Usage

gg_CV(
  x,
  prop = TRUE,
  percent = TRUE,
  drop_zeros = TRUE,
  text = TRUE,
  text_colour = "white",
  text_size = 8,
  text_signif = 2,
  axis_size = 10,
  axis_x_angle = 0
)

## Default S3 method:
gg_CV(
  x,
  prop = TRUE,
  percent = TRUE,
  drop_zeros = TRUE,
  text = TRUE,
  text_colour = "white",
  text_size = 8,
  text_signif = 2,
  axis_size = 10,
  axis_x_angle = 0
)

## S3 method for class 'stat_lda_full'
gg_CV(
  x,
  prop = TRUE,
  percent = TRUE,
  drop_zeros = TRUE,
  text = TRUE,
  text_colour = "white",
  text_size = 8,
  text_signif = 2,
  axis_size = 10,
  axis_x_angle = 0
)

## S3 method for class 'matrix'
gg_CV(
  x,
  prop = TRUE,
  percent = TRUE,
  drop_zeros = TRUE,
  text = TRUE,
  text_colour = "white",
  text_size = 8,
  text_signif = 2,
  axis_size = 10,
  axis_x_angle = 0
)

Arguments

x	confusion matrix or CV_tbl returned by lda
prop	logical if TRUE (default), use accuracies (within an actual class); if FALSE use counts
percent	logical whether to use percentages (default to TRUE)
drop_zeros	logical whether to drop zeros in labels (default to TRUE)
text	logical whether display cell information (default to TRUE)
text_colour	logical whether display cell information (default to white)
text_size	numeric size for cell information, in points (default to 8)
text_signif	integer to feed signif (default to '2“)
axis_size	numeric size for axes labels
axis_x_angle	numeric angle for x-axis labels

Details

when a matrix is passed, it assumes that it is:

• squared

• rows are for actual classes

• columns for predicted classes

• class accuracies are calculated per actual classes, ie rowwise

Examples

# works fine on matrices too
# set.seed(2329) # for the sake of replicability
m <- matrix(sample(1:100), 10)
gg_CV(m)

MomX/Momstats documentation built on April 8, 2020, 8:04 a.m.