A mixture of regression models for multivariate observed variables which contextually involves a dimension reduction step through a linear factor model is proposed. The model estimation is performed via the EM-algorithm and a procedure to compute asymptotic standard errors for the parameter estimates is developed. The proposed approach is applied to the study of students satisfaction towards different aspects of their school as a function of various covariates.
A. Montanari, C. Viroli (2011). Dimensionally reduced mixtures of regression models. JOURNAL OF STATISTICAL PLANNING AND INFERENCE, 141, 1744-1752 [10.1016/j.jspi.2010.11.024].
Dimensionally reduced mixtures of regression models
MONTANARI, ANGELA;VIROLI, CINZIA
2011
Abstract
A mixture of regression models for multivariate observed variables which contextually involves a dimension reduction step through a linear factor model is proposed. The model estimation is performed via the EM-algorithm and a procedure to compute asymptotic standard errors for the parameter estimates is developed. The proposed approach is applied to the study of students satisfaction towards different aspects of their school as a function of various covariates.File in questo prodotto:
Eventuali allegati, non sono esposti
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
