A general trimming approach to robust cluster analysis

  1. García Escudero, Luis Angel
  2. Gordaliza Ramos, Alfonso
  3. Matrán Bea, Carlos
  4. Mayo Iscar, Agustín
Book:
XXX Congreso Nacional de Estadística e Investigación Operativa y de las IV Jornadas de Estadística Pública: actas

Publisher: Comité organizador del XXX Congreso Nacional de Estadística e Investigación Operativa y IV Jornadas de Estadística Pública

ISBN: 978-84-690-7249-3

Year of publication: 2007

Congress: Congreso Nacional de Estadística e Investigación Operativa (30. 2007. Valladolid)

Type: Conference paper

Abstract

We introduce a new method for performing clustering with the aim of ¯tting clusters with di®erent scatters and weights. It is designed by allowing to handle a proportion ® of contaminating data to guarantee the robustness of the method. As a characteristic feature, restrictions on the ratio between the maximum and the minimum eigenvalues of the groups scatter matrices are introduced. This makes the problem to be well-de¯ned and guarantees the consistency of the sample solutions to the population ones. The method covers a wide range of clustering approaches depending on the strength of the chosen restrictions. The proposal includes an algorithm (the TCLUST method) for approximately solving the sample problem.