Current code mixes both spatial and data-based distance metrics. The algorithms actually require all initial processing steps to be conducted with the spatial distances only, and then the tree partitioning to use the data-based distances. That's the reason for current messes like https://github.com/mpadge/spatialcluster/issues/26#issuecomment-1307523058
Current code mixes both spatial and data-based distance metrics. The algorithms actually require all initial processing steps to be conducted with the spatial distances only, and then the tree partitioning to use the data-based distances. That's the reason for current messes like https://github.com/mpadge/spatialcluster/issues/26#issuecomment-1307523058