SEMINAR

MONDAY 26 march at 16.30 in A320 (DINFO, University of Palermo)
Prof. M. A. Dahleh
LIDS Laboratory for Information and Decision Systems
MIT MASSACHUSSETS INSTITUTE OF TECHNOLOGY

will give a seminar entitltled

'' Model Reduction of Hidden Markov Models''

You are all invited

three weeks of control class

After three weeks we know how to find a state space model and
how to find a transfer function, the input/output representation.
We revise some linear algebra and some function analysis, we know how to transform and anti-transform in the s- and z-domain.

This means that we have all the ingredients to answer to the main question we pose to ourselves:
given a system, what is the response of the system to its initial conditions and to a given input.

From the analysis of the response it is possible then to get the main inference on the system itself.

identification and control

A stochastic process has been characterized through its variance and its spectral density.
A stationary process can be seen as the output of a lti system with white noise as imput.
Various process (AR, MA and ARMA) have been studied.