The following pages and posts are tagged with
| Title | Type | Excerpt |
|---|---|---|
| K-means | Page | * The most popular clustering algorithm. * Scales well with many samples. * The algorithm initially defines $$ k $$ points as cluster centroids. $$ k $$ is defined by the user. * It goes through every sample and assigns each sample to the closest... |
| Combining K-means and mean shift | Page | ## Combining mean shift and k-means algorithms to cluster non-convex shapes {% include image.html file="img/kmeans_meanshift.png" caption="from https://jamesxli.blogspot.com/2012/03/on-mean-shift-and-k-means-clustering.html"%} |
| Mean shift | Page | * This algorithm identifies clusters by detecting areas of high density of samples. * The algorithm iteratively shifts each point in the direction of increasing KDE (kernel density estimation) until all points are on a KDE... |