Clustering aims at dividing a data set into groups or clusters that consist of similar data. Fuzzy clustering accepts the fact that the clusters or classes in the data are usually not completely well separated and thus assigns a membership degree between 0 and 1 for each cluster to every datum. We introduce a robust method for fuzzy clustering based on mixtures of Gaussian Factor analyzers. We illustrate our theoretical considerations by simulations and applications to real data. A comparison with probabilistic clustering is also provided
Greselin, F., Garcia-Escudero, L., Mayo-Iscar, A. (2016). A fuzzy version of robust mixtures of Gaussian factor analyzers. In Programme and abstracts. 9th International Conference of the ERCIM (European Research Consortium for Informatics and Mathematics) Working Group on Computational and Methodological Statistics (CMStatistics 2016) (pp. 5-5). Sevilla : Technical Editors: Angela Blanco-Fernandez and Gil Gonzalez-Rodriguez..
A fuzzy version of robust mixtures of Gaussian factor analyzers
Greselin, F;
2016
Abstract
Clustering aims at dividing a data set into groups or clusters that consist of similar data. Fuzzy clustering accepts the fact that the clusters or classes in the data are usually not completely well separated and thus assigns a membership degree between 0 and 1 for each cluster to every datum. We introduce a robust method for fuzzy clustering based on mixtures of Gaussian Factor analyzers. We illustrate our theoretical considerations by simulations and applications to real data. A comparison with probabilistic clustering is also provided
| File | Dimensione | Formato | |
|---|---|---|---|
|
GGM Ercim 2016 A fuzzy version of robust mixtures of Gaussian factor analyzers.pdf
accesso aperto
Dimensione
574.59 kB
Formato
Adobe PDF
|
574.59 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
