python - how are clusters typically retrieved from Mean Shift Clustering? -
i've implemented mean shift clustering in python. used matplotlib , results i'm expecting on clustering on random input. however, i'm confused how 1 reduces cluster each point few clusters.
what mean mean shift outputs each point cluster, though many points may correspond same cluster. had few ways theorized done, seemed odd me not single tutorial looked on explained how extract unique clusters on output:
one have list of clusters each point, , after shifting done each point, check if newly created cluster equal given point (or close distance metric) add original point of shifted point cluster. other wise create new cluster pair [cluster_location: [cluster_points]]
one have hash (dictionary) hash cluster positions , if current cluster position exists, add un shifted point cluster, otherwise create new cluster hash index.
i'm leaning toward doing b, know standard method cluster retrieval mean shift is.
Comments
Post a Comment