Deterministic Annealing Clustering

This software partitions a dataset into clusters of data using the Deterministic Annealing algorithm, a more sophisticated clustering technique than usually available in scientific and commercial software. The results provided are deterministic, i.e. unique (within the machine precision tolerance), whereas results provided by commercially available scientific software usually are not. For instance, results provided by the popular K-means depend on random seeds and therefore necessitate multiple runs to achieve a reasonable convergence. The uniqueness or robustness of the solution is an essential feature for scientific use.