Evaluating the Ambiguity of Fuzzy Clustering with Fuzzy Entropy

In this paper, we strive to interpret the significance of the solution for the categorisation problem by the fuzzy c-means clustering algorithm in terms of the fuzzy membership values obtained by clustering. We use the concept of fuzzy entropy for interpreting the results. A threshold value for fuzzy entropy is computed in our work for labelling sample data as ambiguous or of an uncertain class. Results on images from the Chinese Academy of Sciences- Institute of Automation (CASIA) iris database with features extracted as per Libor Masek's iris segmentation model confirm our argument.