Kernel K-Means Clustering Learner

A LearnerClust for kernel k-means clustering implemented in kernlab::kkmeans(). kernlab::kkmeans() doesn't have a default value for the number of clusters. Therefore, the centers parameter here is set to 2 by default. Kernel parameters have to be passed directly and not by using the kpar list in kkmeans. The predict method finds the nearest center in kernel distance to assign clusters for new data points.

Dictionary

This mlr3::Learner can be instantiated via the dictionary mlr3::mlr_learners or with the associated sugar function mlr3::lrn():

mlr_learners$get("clust.kkmeans")
lrn("clust.kkmeans")

Meta Information

• Task type: “clust”

• Predict Types: “partition”

• Feature Types: “logical”, “integer”, “numeric”

• Required Packages: mlr3, mlr3cluster, kernlab

Parameters

Id	Type	Default	Levels	Range
centers	untyped	-		-
kernel	character	rbfdot	vanilladot, polydot, rbfdot, tanhdot, laplacedot, besseldot, anovadot, splinedot	-
sigma	numeric	-		\([0, \infty)\)
degree	integer	3		\([1, \infty)\)
scale	numeric	1		\([0, \infty)\)
offset	numeric	1		\((-\infty, \infty)\)
order	integer	1		\((-\infty, \infty)\)
alg	character	kkmeans	kkmeans, kerninghan	-
p	numeric	1		\((-\infty, \infty)\)

References

Karatzoglou, Alexandros, Smola, Alexandros, Hornik, Kurt, Zeileis, Achim (2004). “kernlab-an S4 package for kernel methods in R.” Journal of statistical software, 11, 1–20.

Dhillon, S I, Guan, Yuqiang, Kulis, Brian (2004). A unified view of kernel k-means, spectral clustering and graph cuts. Citeseer.

See also

• Chapter in the mlr3book: https://mlr3book.mlr-org.com/chapters/chapter2/data_and_basic_modeling.html#sec-learners

• Package mlr3extralearners for more learners.

• Dictionary of Learners: mlr3::mlr_learners

• as.data.table(mlr_learners) for a table of available Learners in the running session (depending on the loaded packages).

• mlr3pipelines to combine learners with pre- and postprocessing steps.

• Extension packages for additional task types:

  • mlr3proba for probabilistic supervised regression and survival analysis.

  • mlr3cluster for unsupervised clustering.

• mlr3tuning for tuning of hyperparameters, mlr3tuningspaces for established default tuning spaces.

Other Learner: mlr_learners_clust.MBatchKMeans, mlr_learners_clust.SimpleKMeans, mlr_learners_clust.agnes, mlr_learners_clust.ap, mlr_learners_clust.bico, mlr_learners_clust.birch, mlr_learners_clust.cmeans, mlr_learners_clust.cobweb, mlr_learners_clust.dbscan, mlr_learners_clust.dbscan_fpc, mlr_learners_clust.diana, mlr_learners_clust.em, mlr_learners_clust.fanny, mlr_learners_clust.featureless, mlr_learners_clust.ff, mlr_learners_clust.hclust, mlr_learners_clust.hdbscan, mlr_learners_clust.kmeans, mlr_learners_clust.mclust, mlr_learners_clust.meanshift, mlr_learners_clust.optics, mlr_learners_clust.pam, mlr_learners_clust.xmeans

Super classes

mlr3::Learner -> mlr3cluster::LearnerClust -> LearnerClustKKMeans

Methods

Public methods

• LearnerClustKKMeans$new()

• LearnerClustKKMeans$clone()

Inherited methods

• mlr3::Learner$base_learner()
• mlr3::Learner$configure()
• mlr3::Learner$encapsulate()
• mlr3::Learner$format()
• mlr3::Learner$help()
• mlr3::Learner$predict()
• mlr3::Learner$predict_newdata()
• mlr3::Learner$print()
• mlr3::Learner$selected_features()
• mlr3::Learner$train()
• mlr3cluster::LearnerClust$reset()

---

Method new()

Creates a new instance of this R6 class.

Usage

LearnerClustKKMeans$new()

---

Method clone()

The objects of this class are cloneable with this method.

Usage

LearnerClustKKMeans$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# Define the Learner and set parameter values
learner = lrn("clust.kkmeans")
print(learner)
#> 
#> ── <LearnerClustKKMeans> (clust.kkmeans): Kernel K-Means ───────────────────────
#> • Model: -
#> • Parameters: centers=2
#> • Packages: mlr3, mlr3cluster, and kernlab
#> • Predict Types: [partition]
#> • Feature Types: logical, integer, and numeric
#> • Encapsulation: none (fallback: -)
#> • Properties: complete, exclusive, and partitional
#> • Other settings: use_weights = 'error'

# Define a Task
task = tsk("usarrests")

# Train the learner on the task
learner$train(task)
#> Using automatic sigma estimation (sigest) for RBF or laplace kernel 

# Print the model
print(learner$model)
#> Spectral Clustering object of class "specc" 
#> 
#>  Cluster memberships: 
#>  
#> 1 1 1 1 1 1 2 1 1 1 2 2 1 2 2 2 2 1 2 1 1 1 2 1 1 2 2 1 2 1 1 1 1 2 2 1 1 2 1 1 2 1 1 2 2 1 1 2 2 1 
#>  
#> Gaussian Radial Basis kernel function. 
#>  Hyperparameter : sigma =  0.00043279660093855 
#> 
#> Centers:  
#>          [,1]     [,2]     [,3]  [,4]
#> [1,] 226.2333 10.13333 25.79333 69.40
#> [2,]  87.5500  4.27000 14.39000 59.75
#> 
#> Cluster size:  
#> [1] 30 20
#> 
#> Within-cluster sum of squares:  
#> [1] 1799973.6  213075.9
#> 

# Make predictions for the task
prediction = learner$predict(task)

# Score the predictions
prediction$score(task = task)
#> clust.dunn 
#> 0.04322918