A 12×12 grid of neurons. Each sample pulls the nearest neuron and its grid neighbours toward it, and the neighbourhood and learning rate shrink over time.
Tap to change the data shape. The same ?seed= always gives the same run.
LITKohonen’s self-organising map (1982). Every neuron must stay inside the data space, or the page throws.
AMBERNeighbour order preserved by the sheet is visible, not proved.