A LOCAL GEOMETRICAL PROPERTIES APPLICATION TO FUZZY CLUSTERING

Possibilistic clustering is seen increasingly as a suitable means to r esolve the limitations resulting from the constraints imposed in the f uzzy C-means algorithm. Studying the metric derived from the covarianc e matrix we obtain a membership function and an objective function whe ther the Mahalanobis distance or the Euclidean distance is used. Apply ing the theoretical results using the Euclidean distance we obtain a n ew algorithm called fuzzy-minimals, which detects the possible prototy pes of the groups of a sample. We illustrate the new algorithm with se veral examples. (C) 1998 Elsevier Science B.V. All rights reserved.