ml_kmeans

ml_kmeans(x, k, seed=None, kmeans_plus_plus=True, max_iterations=1000)

k-means clustering.

Mirrors the C# KMeans class: partitions the rows of x into k clusters, each row belonging to the cluster with the nearest centroid. Initialization is k-means++ by default.

Two behaviours inherited from upstream are worth knowing:

• labels are 0-BASED in both Python and R, matching the library’s own indexing (the same choice :func:shortest_path made for node indices).
• With k = 1 the algorithm stops before its first update step, so the single reported “cluster mean” is a randomly chosen observation rather than the mean of x.

Parameters

Name	Type	Description	Default
x	array_like	A 2-D array with one row per observation, or a 1-D array for a single feature.	required
k	int	The number of clusters.	required
seed	int	PRNG seed. None (the default) uses the computer clock, so the fit is not reproducible; supply a seed for a reproducible fit.	None
kmeans_plus_plus	bool	Use k-means++ initialization rather than a uniform draw.	True
max_iterations	int	The iteration cap.	1000

Returns

Name	Type	Description
	dict	means (a k by x.shape[1] array of centroids), labels (a 0-based integer array, one per row of x), and iterations.

Examples

>>> from corehydropy import ml_kmeans
>>> x = [[1, 2], [1.2, 2.1], [0.8, 1.9], [8, 9], [8.3, 9.2], [7.9, 8.8]]
>>> fit = ml_kmeans(x, k=2, seed=12345)
>>> fit["means"].shape
(2, 2)