Elena Stanghellini
Dipartimento di Scienze Statistiche
Universita' di Perugia, Italy
Via A.Pascoli, 1 C.P. 1315 Succ.1
06100 Perugia
E-mail: stanghel@stat.unipg.it
Phone: +39-075-585 5229
Fax: +39 -075- 585 43242
Joe Whittaker
Department of Mathematics and Statistics
Lancaster University, Fylde College, Lancaster, UK, LA1 4YF
E-mail: Joe Whittaker
Phone: +44-1524-593 960
Fax: +44-1524-592 681
We propose a chain graph with unobserved variables to model a multivariate time series. We assume that an underlying common trend linearly affects the observed time series, but we do not restrict our analysis to models where the underlying factor accounts for all the contemporary correlations of the series. The residual correlation is modelled using results of graphical models. Modelling the associations left unexplained is an alternative to augmenting the dimension of the underlying factor. It is justified when a clear interpretation of the residual associations is available. It is also an informative way to explore sources of deviation from standard dynamic single-factor models.
chain graph, dynamic factor analysis, hidden variables, identifiabilty.
email to the first author