R. Kuusik

The Super-Fast Algorithm of Hierarchical Clustering
and The Theory of Monotonic Systems (pdf)

A new hierarchical clustering algorithm based on theory of monotonic systems is described. Let us have N*M data table (N is number of objects, M is number of attributes), where each attribute j, j=1,...,M has a discrete value hj = 0,1,...,K-1. If N=K^M, then the complexity of described algorithm is O(N*N=N^2) for extracting all (K+1)^M clusters.