KMEANS: KMEANS

In ChrisKeefe/UnsupLP1: A sample package exposing some functions for Unspervised Learning

Description

Generates a K-means clustering of a given feature matrix, with K cluster centers. K-means is a common non-probabilistic clustering method, useful when data can be fit to circular clusters.

Usage

KMEANS(data_matrix, K)

Arguments

data_matrix	An N x M matrix, containing N rows (observations) and M columns (data features)
K	A positive integer describing the number of clusters the clustering should produce.

Value

A list containing the following results:

cluster	a vector of length N containing the cluster ids of each observation. Cluster ids are positive integers from 1:K inclusive.
centers	a matrix of cluster centers, with K rows, and M columns.
withinss	a vector of length K, containing within-cluster sums of squares for each cluster.
tot.withinss	the total within-cluster sum of squares
size	a vector of length K, containing the number of points in each cluster.

See Also

stats::kmeans for a more robust implementation.

Examples

# Cluster the iris data set into three clusters:
KMEANS(iris[1:4], K=3)

ChrisKeefe/UnsupLP1 documentation built on Oct. 8, 2020, 5:37 a.m.