Robust fuzzy clustering algorithms

A class of fuzzy clustering algorithms based on a recently introduced noise cluster concept is proposed. A noise prototype is defined such that it is equidistant to all the points in the data set. This allows detection of clusters among data with or without noise. It is shown that this concept is applicable to all the generalizations of fuzzy and hard K-means algorithms. Various applications are considered. The application of this concept to a variety of regression problems is also considered. It is shown that the results of this approach are comparable to those of many robust regression techniques