Plot for Hierarchical Clustering Learners

Generates plots for hierarchical clusterers, depending on argument type:

• "dend" (default): dendrograms using factoextra package.

• "scree": scree plot that shows the number of possible clusters on x-axis and the height on the y-axis.

Note that learner-specific plots are experimental and subject to change.

Usage

# S3 method for LearnerClustHierarchical
autoplot(object, type = "dend", ...)

Arguments

object

(mlr3cluster::LearnerClustAgnes | mlr3cluster::LearnerClustDiana | mlr3cluster::LearnerClustHclust).

type

(character(1)):
Type of the plot. See description.

...

(any): Additional arguments, passed down to function factoextra::fviz_dend() in package factoextra.

Value

ggplot2::ggplot() object.

Theme

The theme_mlr3() and viridis color maps are applied by default to all autoplot() methods. To change this behavior set options(mlr3.theme = FALSE).

Examples

library(mlr3)
library(mlr3cluster)
library(mlr3viz)

task = mlr_tasks$get("usarrests")

# agnes clustering
learner = mlr_learners$get("clust.agnes")
learner$train(task)
autoplot(learner)
#> Warning: `guides(<scale> = FALSE)` is deprecated. Please use `guides(<scale> = "none")` instead.


# diana clustering
learner = mlr_learners$get("clust.diana")
learner$train(task)
autoplot(learner,
  k = learner$param_set$values$k, rect_fill = TRUE,
  rect = TRUE, rect_border = "red")
#> Warning: `guides(<scale> = FALSE)` is deprecated. Please use `guides(<scale> = "none")` instead.


# hclust clustering
learner = mlr_learners$get("clust.hclust")
learner$train(task)
autoplot(learner, type = "scree")