On the Selection of Parameter m in Fuzzy c-Means: A Computational Approach

Several clustering algorithms include one or more parameters to be fixed before its application. This is also the case of fuzzy c-means, one of the most well-known fuzzy clustering algorithms, where two parameters c and m are required. c corresponds to the number of clusters and m to the fuzziness of the solutions. The selection of these parameters is a critical issue because a bad selection can blur the clusters in the data. In this paper we propose a method for selecting an appropriate parameter m for fuzzy c-means based on an extensive computation. Our approach is based on the application of the clustering algorithm to several instantiations of the same data with different degrees of noise