ScratchData LogoScratchData
Back to Nishpish's profile

Point Clusterer (k-means)

NINishpish•Created July 21, 2020
Point Clusterer (k-means)
63
48
640 views
View on Scratch

Instructions

Click on the screen to create a point. Make up to 5 clusters out of these points. Then press calculate. It will take around 10 seconds. The program attempts to find the clusters you drew. You can then manually adjust how many clusters there are if it doesn’t seem right. Enjoy!

Description

Uses k-means clustering. First, it clusters the points with random starting values. It repeats a bunch of times with different starting values, and selects the best result. The entire process is repeated for a few different values of k, in an attempt to select the best value. K-means is BAD.

Project Details

Project ID412966246
CreatedJuly 21, 2020
Last ModifiedJune 21, 2021
SharedJuly 21, 2020
Visibilityvisible
CommentsAllowed