On the mathematical and numerical properties of the fuzzy \(c\)-means algorithm (Q1198834)

The fuzzy \(c\)-means algorithm is a valid method for fuzzy clustering proposed by \textit{J. C. Bezdek} [see J. Cybernetics 3, No. 3, 58-73 (1973; Zbl 0294.68035)]. Then it has been developed further by some authors. The present authors investigate some numerical properties of the fuzzy \(c\)- means algorithm, which are related to the points generated in the course of using the fuzzy \(c\)-means algorithm by means of the idea of reduced objective functions. Seven quantities for stopping the algorithm are investigated, and it is empirically shown that these quantities converge linearly. I feel that the empirical result is very interesting (section 4.3: empirical study of the order of convergence).