this feature fixes up subsets at the end of an rcluster run. This only happens if the user specifies the —all-states or —min-subset-size flags. At the end of the rcluster run, any subsets that fail these tests (i.e. don’t have all states and/or are smaller than the specified minimum size) are merged with their nearest neighbour subsets. Nearest neighbour is defined by the manhattan distance, using the user-specified weights.
this feature fixes up subsets at the end of an rcluster run. This only happens if the user specifies the —all-states or —min-subset-size flags. At the end of the rcluster run, any subsets that fail these tests (i.e. don’t have all states and/or are smaller than the specified minimum size) are merged with their nearest neighbour subsets. Nearest neighbour is defined by the manhattan distance, using the user-specified weights.