Approximate Determination of q-Parameter for FCM with Tsallis Entropy Maximization

抄録

This paper considers a fuzzy c-means (FCM) clustering algorithm in combination with deterministic annealing and the Tsallis entropy maximization. The Tsallis entropy is a q-parameter extension of the Shannon entropy. By maximizing the Tsallis entropy within the framework of FCM, statistical mechanical membership functions can be derived. One of the major considerations when using this method is how to determine appropriate values for q and the highest annealing temperature, T_high, for a given data set. Accordingly, in this paper, a method for determining these values simultaneously without introducing any additional parameters is presented, where the membership function is approximated using a series expansion method. The results of experiments indicate that the proposed method is effective, and both q and T_high can be determined automatically and algebraically from a given data set.