/MeanShift

Custom Mean Shift algorithm (mostly for educational purposes)

Primary LanguagePython

custom mean shift algorithm. Used Sentdex's great tutorial video series as guidance (see below for links) but tried to make my own version that I think might run faster as I do not make a large iterable just to take the average vector when calculating the new average vector by weighting vectors closer to the current cluster.

Note: no gurantees on how consistent it works but eyeballed a handful of times using sklearn blobs and seems to do well when checking it visually on a 2-d dataset.

Sentdex references: Videos: https://www.youtube.com/watch?v=P-iAd8b7zl4&list=PLQVvvaa0QuDfKTOs3Keq_kaG2P55YRn5v&index=41 https://www.youtube.com/watch?v=k1alPDpSGBE&list=PLQVvvaa0QuDfKTOs3Keq_kaG2P55YRn5v&index=42

Website: https://pythonprogramming.net/weighted-bandwidth-mean-shift-machine-learning-tutorial/