Mixtures of Gaussian factors are powerful tools for modeling an unobserved heterogeneous population, offering at the same time dimension reduction and model-based clustering. Unfortunately, the high prevalence of spurious solutions and the disturbing effects of outlying observations, along maximum likelihood estimation, open serious issues. We consider restrictions for the component covariances, to avoid spurious solutions, and trimming, to provide robustness against violations of normality assumptions of the underlying latent factors. A detailed AECM algorithm for this new approach is presented. Simulation results and an application to the AIS dataset show the aim and effectiveness of the proposed methodology
Garcìa Escudero, L., Gordaliza, A., Greselin, F., Ingrassia, S., & Mayo Iscar, A. (2015). Robust estimation for mixtures of Gaussian factor analyzers. In I. Gijbels, M. Hubert, B.U. Park, & R. Welsch (a cura di), Programme and Abstracts 8th International Conference of the ERCIM WG on Computational and Methodological Statistics (CMStatistics 2015), London (UK), December 12-14, 2015 (pp. 163-163). London : CFE and CMStatistics networks.
Citazione: | Garcìa Escudero, L., Gordaliza, A., Greselin, F., Ingrassia, S., & Mayo Iscar, A. (2015). Robust estimation for mixtures of Gaussian factor analyzers. In I. Gijbels, M. Hubert, B.U. Park, & R. Welsch (a cura di), Programme and Abstracts 8th International Conference of the ERCIM WG on Computational and Methodological Statistics (CMStatistics 2015), London (UK), December 12-14, 2015 (pp. 163-163). London : CFE and CMStatistics networks. |
Titolo: | Robust estimation for mixtures of Gaussian factor analyzers |
Autori: | Garcìa Escudero, L; Gordaliza, A; Greselin, F; Ingrassia, S; Mayo Iscar, A |
Autori: | |
Presenza di un coautore afferente ad Istituzioni straniere: | Si |
Tipo: | Capitolo o saggio |
Carattere della pubblicazione: | Scientifica |
Data di pubblicazione: | 2015 |
Lingua: | English |
Titolo del libro: | Programme and Abstracts 8th International Conference of the ERCIM WG on Computational and Methodological Statistics (CMStatistics 2015), London (UK), December 12-14, 2015 |
ISBN: | 978-9963-2227-0-4 |
Appare nelle tipologie: | 03 - Contributo in libro |
File in questo prodotto:
File | Descrizione | Tipologia | Licenza | |
---|---|---|---|---|
GGGIM Ercim 2015 Robust estimation for mixtures of Gaussian factor analyzers.pdf | N/A | Open Access Visualizza/Apri |