Timeline for Does the plane clustered to minimize sum distances^2 to clusters centers ( inertia / "K-means") produce hexagonal clusters / hexagonal lattice?
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 14, 2020 at 21:01 | vote | accept | Alexander Chervov | ||
| Jul 14, 2020 at 8:27 | answer | added | Adam P. Goucher | timeline score: 3 | |
| Jun 4, 2020 at 9:52 | comment | added | Alexander Chervov | Not quite. Of course, points are clustered as most close to cluster centers, so it is, of course, Voronoi diagram for cluster centers. But, the main point of the question, is that clusters centers behave not randomly , but seems to tend to hexagonal lattice. Please pay attention, that we are optimizing clusters centers positions, they are not fixed. | |
| Jun 4, 2020 at 3:40 | comment | added | Sandeep Silwal | Since a point p is assigned to center c iff c is the closest center to p otherwise the cost decreases, cluster centers partition the space into voronoi diagrams (en.wikipedia.org/wiki/Voronoi_diagram). So your question is more about how voronoi diagrams of random points look like | |
| Jun 3, 2020 at 22:48 | history | edited | Alexander Chervov | CC BY-SA 4.0 |
added 71 characters in body
|
| Jun 3, 2020 at 20:32 | history | asked | Alexander Chervov | CC BY-SA 4.0 |