Interval type-2 fuzzy clustering for membership function generation

This paper presents the basic theory of the Fuzzy C-Means (FCM) algorithm, as well as the proposed IT2 FCM algorithm, which is an extension of the FCM algorithm, that implements techniques of type-2 fuzzy sets, this in order to improve fuzzy data clustering, being able to handle this algorithm with higher degree of uncertainty and be less prone to noise. The approach is illustrated with plots of clusters generated by the IT2 FCM algorithm and memberships functions of type-2, this was done to observe if the Type-2 membership functions generated by the membership matrices produced by the IT2 FCM algorithm for lower and upper limits of the range, present a significant footprint of uncertainty.