A reweighting approach to robust clustering
- 304 Downloads
An iteratively reweighted approach for robust clustering is presented in this work. The method is initialized with a very robust clustering partition based on an high trimming level. The initial partition is then refined to reduce the number of wrongly discarded observations and substantially increase efficiency. Simulation studies and real data examples indicate that the final clustering solution has both good properties in terms of robustness and efficiency and naturally adapts to the true underlying contamination level.
KeywordsCluster analysis Trimming Robustness Minimum covariance determinant estimator
The research was partially supported by the Spanish Ministerio de Economía y Competitividad y fondos FEDER, Grant MTM2014-56235-C2-1-P, and by Consejería de Educación de la Junta de Castilla y León, Grant VA212U13. We are grateful to Gallup, Inc. and the Voices of the Hungry project, FAO, for access to the GWP/FIES data. We also would like to thank the AE and two referees for kind comments.
- Ballard, T.J., Kepple, A.W., Cafiero, C.: The food insecurity experience scale: developing a global standard for monitoring hunger worldwide. Technical report, Food and Agriculture Organization of the United Nations, Rome (2013)Google Scholar
- Cafiero, C., Nord, M., Viviani, S., del Grossi, M.E., Ballard, T.J., Kepple, A.W., Miller, M., Nwosu, C.: Methods for estimating comparable rates of food insecurity experienced by adults throughout the world. Technical report, Food and Agriculture Organization of the United Nations, Rome (2016)Google Scholar
- Gallup: Worldwide Research Methodology and Codebook. Gallup Inc, Washington (2015)Google Scholar
- Hennig, C.: Dissolution point and isolation robustness: robustness criteria for general cluster analysis methods. J. Multivar. Anal. 99, 1154–1176 (2008)Google Scholar