**Spherical k-means** is an unsupervised clustering algorithm where the lengths of all vectors being compared are normalized to 1, so that they differ in direction but not in magnitude. Clustering can then be carried out more efficiently by measuring the angles between the vectors (**cosine similarity**) than by using the standard k-means algorithm.

Spherical k-means is preferred to standard k-means:

- when the magnitude of the vectors is irrelevant in terms of what the data represents;
- when the magnitude of the vectors is not particularly important in terms of what the data represents and the vectors have a large number of dimensions, because spherical k-means is a more efficient learning technique.

- alias
- Cosine similarity
- subtype
- has functional building block
- FBB_Classification
- has input data type
- IDT_Vector of quantitative variables
- has internal model
- has output data type
- ODT_Vector of quantitative variables ODT_Classification
- has learning style
- LST_Unsupervised
- has parametricity
- PRM_Nonparametric
- has relevance
- REL_Relevant
- uses
- sometimes supports
- mathematically similar to