An Optimal Choice Method of Parameter m in FCM Clustering Algorithm

Fuzzy c-mean(FCM) is a popular and effective algorithm for obtaining an optimal fuzzy (partition) of data set by minimizing an objective function, in which the weighting exponent m plays an important role. To perform the FCM algorithm, a proper value should be assigned to the parameter m. It thus is essential for us to study the optimal choice of m. For this end, this paper presents an (optimal) choice method for m based on fuzzy decision theory. The experimental results illustrate the (effectiveness) of the proposed method.