Divisive Methods (of Cluster Analysis):
In divisive methods of hierarchical cluster analysis , the clusters obtained at the previous step are subdivided into smaller clusters. Such methods start from a single cluster comprising of all N objects, and, after N-1 steps, they end with N clusters comprising a single object. The criterion of such subdivision is based on the distance between two new clusters, as specified by the chosen linkage function
A complementary concept is agglomerative methods .