Currently for hierarchical agglomerative clustering, we can only retrieve the clusters based on the total number of clusters. One of the features of this bottom-to-top clustering is we can stop the clustering process (or cut the clustering tree) at any distance among clusters.
If we compare to Scipy implementation for creating flat clusters, the current implementation supports criterion='maxclust' only. This pull request make criterion='distance' can be done.
Currently for hierarchical agglomerative clustering, we can only retrieve the clusters based on the total number of clusters. One of the features of this bottom-to-top clustering is we can stop the clustering process (or cut the clustering tree) at any distance among clusters.
If we compare to Scipy implementation for creating flat clusters, the current implementation supports
criterion='maxclust'only. This pull request makecriterion='distance'can be done.