RT info:eu-repo/semantics/article
T1 Robust Constrained Fuzzy Clustering
A1 Fritz, Heinrich
A1 García Escudero, Luis Ángel
A1 Mayo Iscar, Agustín
K1 Estádistica
AB It is well-known that outliers and noisy data can be very harmful when applyingclustering methods. Several fuzzy clustering methods which are ableto handle the presence of noise have been proposed. In this work, we proposea robust clustering approach called F-TCLUST based on an “impartial”(i.e., self-determined by data) trimming. The proposed approach considersan eigenvalue ratio constraint that makes it a mathematically well-definedproblem and serves to control the allowed differences among cluster scatters.A computationally feasible algorithm is proposed for its practical implementation.Some guidelines about how to choose the parameters controlling theperformance of the fuzzy clustering procedure are also given.
YR 2013
FD 2013
LK http://uvadoc.uva.es/handle/10324/21850
UL http://uvadoc.uva.es/handle/10324/21850
LA spa
NO Information Sciences, 245, 38-52.
DS UVaDOC
RD 25-abr-2024