RT info:eu-repo/semantics/preprint
T1 A fast algorithm for robust constrained clustering
A1 García Escudero, Luis Ángel
A1 Fritz, Heinrich
A1 Mayo Iscar, Agustín
A2 Universidad de Valladolid
K1 Statistics
AB The application of “concentration” steps is the main principle behind Forgy’sk-means algorithm and Rousseeuw and van Driessen’s fast-MCD algorithm.Despite this coincidence, it is not completely straightforward to combine bothalgorithms for developing a clustering method which is not severely affectedby few outlying observations and being able to cope with non spherical clusters.A sensible way of combining them relies on controlling the relative clusterscatters through constrained concentration steps. With this idea in mind,a new algorithm for the TCLUST robust clustering procedure is proposedwhich implements such constrained concentration steps in a computationallyefficient fashion.
YR 2013
FD 2013
LK http://uvadoc.uva.es/handle/10324/5964
UL http://uvadoc.uva.es/handle/10324/5964
LA eng
NO Estadística e IO
DS UVaDOC
RD 21-nov-2024