Dynamic Clustering Visualiser

Speed faster

Data parameters

Data geometry

Points 240

Noise 0.04

Spread 1.00

Seed

Clustering parameters

Method

Clusters / components 3

Fuzziness 2.00

eps 0.17

minPts 5

Min cluster size 16

Bandwidth 0.23

Linkage

Cut clusters 3

Graph neighbours 8

Show method mechanics

Show weak membership where available

k-means

Choose k centres. Assign each point to its nearest centre. Move each centre to the mean of the points assigned to it. Repeat until stable.

Current step

Press Step: each point chooses its nearest centre.

Model and methodology

Model

The selected algorithm groups a generated two-dimensional dataset according to its own distance, density or connectivity criterion. The step controls expose intermediate assignments; the algorithm-specific explanation describes the current procedure.

Assumptions and limitations

These synthetic datasets illustrate algorithm behaviour. A cluster is not necessarily a natural category: its shape and membership depend on the metric, parameters, initialisation and data geometry. Compare algorithms on the same dataset and inspect failed as well as successful partitions. Displayed metrics summarise the chosen criterion; they do not establish scientific validity.

References and implementation

See the source implementation and scikit-learn clustering guide for algorithm definitions and their assumptions. This visualiser implements its own educational versions.