This question pertains to inference in multiply connected networks
representing dynamic processes (complex DBNs).
Is there anyone within our community that is, or has, implimented
inference in such networks with continuous variables.
The network I am dealing with does not easily lend itself to a Linear
Dynamical System representation (which could be solved with a Kalman
filter/smoother). I have already been down this path.
I am particularly interested in the merits of any other inference
techniques applied to these networks - particularly large networks. If
there is anyone that has successfully implimented such networks, I would
appreciate either some references for further reading or some direct
discussion of the issues that arised from the algorithm(s) used.
Thanks in advance,
Tim Wilkin
--
____________________________________________________
[EMAIL PROTECTED] *
Tim Wilkin Ph: +61 3 9905 9677 * .*
Aerosonde Project Fax: +61 3 9905 9689
_____________________________________________*_______
CRC for Southern Hemisphere Meteorology, Monash Univ
School of Computer Sci. & Software Eng., Monash Univ
Department of Mathematics & Statistics, Monash Univ
----------------------------------------------------
