Optimization of Parametrized Divergences in Fuzzy c-Means.
Tina Geweniger, Marika Kästner, Thomas Villmann · 2011
Abstract. Wepropose theutilization ofdivergences as dissimilarity measure in the Fuzzy c-Means algorithm for the clustering of functional data. Further we adapt the relevance parameter to improve the data representation and therefore obtain more accurate clusterings in terms of separation and compactness. We show for two example applications that this method leads to improved performance. 1