Bayes Nets of Time Series: Stochastic Realizations and Projections
Graphical models in which every node holds a time-series are developed using special conditions from static multivariate Gaussian processes, particularly the notion of lattice conditional independence (LCI), due to Anderson and Perlman (1993). Under certain “feedback free” conditions, LCI imposes a special zero structure on the state space representation of processes which have a stochastic realisation. This structure comes directly from the transitive directed acyclic graph (TDAG) which is in one-to-one correspondence with the Boolean Hilbert lattice of the LCO formulation. Simple AR(1) examples are presented.
KeywordsConditional Independence Multivariate Normal Distribution Gaussian Case Boolean Lattice Conditional Covariance
Unable to display preview. Download preview PDF.
- Caines, P., Deardon, R., and Wynn, H. (2003). Conditional orthogonality and conditional stochastic realization. Directions in Mathematical Systems Theory and Optimization, Lecture Notes in Control and Information Sciences, A. Rantzer and C.I. Byrnes (Eds.), 286, 71-84.Google Scholar
- Kadison, R. and Ringrose, J. (1983). Fundamentals of the Theory of Operator Algebras, volume 1. . Volume 15 of Graduate Studies in MathematicsAmerican Mathematical Society, San Diego.Google Scholar
- Lauritzen, L. (1996). Graphical Models. Clarendon Press, Oxford.Google Scholar