cluster

Documentation

Overview

Package cluster provides interfaces and types for data clustering in ℝⁿ.

Index

• type Center
• type Clusterer
• type Indices
• type Interface
• type Point
• type Value
• type Weighter

Types

type Center

type Center interface {
	Point

	// Members returns a set of indices into the slice returned by Clusterer.Values() that
	// refers to the Values associated with the Center.
	Members() Indices
}

A Center is a representation of a cluster center.

type Clusterer

type Clusterer interface {
	// Cluster the data.
	Cluster() error

	// Centers returns a slice of centers of the clusters.
	Centers() []Center

	// Values returns the internal representation of the original data.
	Values() []Value
}

Clusterer is the common interface implemented by clustering types.

type Indices

type Indices []int

Indices is a list of indexes into a array or slice of Values.

type Interface

type Interface interface {
	Len() int               // Return the length of the data vector.
	Values(i int) []float64 // Return the data values for element i as a slice of float64.
}

Interface is a type that can be clustered by a Clusterer.

type Point

type Point interface {
	V() []float64
}

Point represents a point in ℝⁿ.

type Value

type Value interface {
	Point

	// Cluster returns an index into the slice returned by Clusterer.Centers() that
	// refers to the Center associated with the Value.
	Cluster() int
}

A Value is the representation of a data point within the clustering object.

type Weighter

type Weighter interface {
	Weight(i int) float64 // Return the weight for element i.
}

Weighter is an extension of the Interface that allows values represented by the Interface to be differentially weighted.