KMeans class

A class for organizing k-means computations with utility methods for finding a good solution.

https://en.wikipedia.org/wiki/K-means_clustering

Uses k-means++ (https://en.wikipedia.org/wiki/K-means%2B%2B) for the initial means, and then iterates.

The points accepted by KMeans can have any number of dimensions and are represented by List<double>. The distance function is hard-coded to be the euclidan distance.

Constructors

KMeans(List<List<double>> points, {List<int> ignoredDims = _emptyList, int labelDim = -1})

points gives the list of points to cluster. ignoreDims gives a list of dimensions of points to ignore. labelDim gives the index of a label in each point if there is one.

Properties

hashCode → int

The hash code for this object.

no setter, inherited

ignoredDims → List<int>

Dimensions of points that are ignored.

final

labelDim → int

The index of the label. Defaults to -1 if the data is unlabeled.

final

points → List<List<double>>

The points to cluster.

final

runtimeType → Type

A representation of the runtime type of the object.

no setter, inherited

Methods

bestFit({int maxIterations = 300, int seed = 42, int minK = 2, int maxK = 20, int trialsPerK = 1, KMeansInitializer init = KMeansInitializers.kMeansPlusPlus, double tolerance = defaultPrecision, bool useExactSilhouette = false}) → Clusters

Finds the 'best' k-means Cluster over a range of possible values of 'k'.

fit(int k, {int maxIterations = 300, int seed = 42, KMeansInitializer init = KMeansInitializers.kMeansPlusPlus, double tolerance = defaultPrecision}) → Clusters

Returns a Cluster of points into k clusters.

noSuchMethod(Invocation invocation) → dynamic

Invoked when a nonexistent method or property is accessed.

inherited

toString() → String

A string representation of this object.

inherited

Operators

operator ==(Object other) → bool

The equality operator.

inherited

Constants

defaultPrecision → const double

Numbers closer together than this are considered equivalent.