GRAPHICAL MODELLING OF MULTIVARIATE TIME SERIES WITH LATENT VARIABLES

Explore this paper's citation graph

Summary

This paper describes a new graphical approach for modelling the dependence structure of mul- tivariate stationary time series that are afiected by latent variables based on mixed graphs, and shows that these models can be viewed as graphical ARMA models that satisfy the Granger causality restrictions encoded by general mixed graphs.

Type
article
Published
2006-01-01
Cited by
18
References
34

Keywords

Spurious relationship, Latent variable, Graphical model, Autoregressive model, Granger causality

References

Cited by

Related papers