On Fuzzy Clustering Algorithms for Nominal Data

This paper presents three fuzzy clustering algorithms for nominal data. The first algorithm is similar to a conventional algorithm for vectorial data developed by introducing variables for controlling the cluster size. The second algorithm is similar to a conventional algorithm for vectorial data developed by regularizing another conventional algorithm for vectorial data with Kullback-Leibler divergence. The third algorithm is developed by regularizing the first algorithm mentioned above with q-divergence. Finally, some numerical experiments are conducted to investigate the features of the proposed algorithms.