WeightBoxplot {sharp}

Stable attribute weights

Description

Creates a boxplots of the distribution of (calibrated) median attribute weights obtained from the COSA algorithm across the subsampling iterations. See examples in Clustering.

Usage

WeightBoxplot(
  stability,
  at = NULL,
  argmax_id = NULL,
  col = NULL,
  boxwex = 0.3,
  xlab = "",
  ylab = "Weight",
  cex.lab = 1.5,
  las = 3,
  frame = "F",
  add = FALSE,
  ...
)

Arguments

stability

output of Clustering.

at

coordinates along the x-axis (more details in boxplot).

argmax_id

optional indices of hyper-parameters. If argmax_id=NULL, the calibrated hyper-parameters are used.

col

optional vector of colours.

boxwex

box width (more details in boxplot).

xlab

label of the x-axis.

ylab

label of the y-axis.

cex.lab

font size for labels.

las

orientation of labels on the x-axis (see par).

frame

logical indicating if the box around the plot should be drawn (more details in boxplot).

add

logical indicating if the boxplot should be added to the current plot.

...

additional parameters passed to boxplot).

Value

A boxplot.

See Also

Clustering