Machine Learning

k-Means

Pick k centers, colour each point by its nearest one, move each center to the mean of its points, and repeat.

Click the canvas to add a point

Inertia J—

—

Clusters k 3

Random data

iteration—

next step—

points—

status—

What to observe

Press Step a few times. The two halves of the loop take turns: each center slides to the mean of its color, then every point is recolored to its nearest center.
Hit Reset a few times on the same points. k-means can settle intodifferent groupings, because it finds a local minimum of J, not the global one.
On Moons k-means insists on round, straight-edged clusters and cuts each crescent in half: the blind spot DBSCAN fixes.

Shortcuts:space run/pause · s step · r reset · f fullscreen