To the best of our knowledge, alternatives to k-means such as the previous have
not been applied to RS.
The simplicity and efficiency of the k-means algorithm shadows possible alternatives. It is not clear whether density-based or hierarchicalclustering approaches have anything to offer in the RS arena.
On the other hand,
message-passing algorithms have been shown to be more efficient and their graphbased paradigm can be easilytranslated to the RS problem. It is possible that we seeapplications of these algorithms in the coming years.