Linear statistical models and stochastic realization theory

  • Lorenzo Finesso
  • Giorgio Picci
Session 8 Identification And Detection
Part of the Lecture Notes in Control and Information Sciences book series (LNCIS, volume 62)


The problem of representing a given gaussian zero mean random vector y by linear statistical models is considered. This is a concrete formulation of a simple stochastic realization problem. Let y=[y′1],y′2]′ be any partition of y into two disjoint subvectors y1, y2. It is shown that to every random vector x, making y1 and y2 conditionally independent given x there corresponds an (essentially unique) model of y of the form
$$\begin{gathered}y_1 = H_1 x + n_1 \hfill \\y_2 = H_2 x + n_2 \hfill \\\end{gathered} $$
where H1 and H2 are deterministic matrices, n1 and n2 are mutually independent noise terms and each ni(i=1,2) is independent of x. The family of all realizations of y of the form (0) is analyzed both probabilistically and from the point of view of explicit computation of the parameters. Possible applications especially to the theory of Factor Analysis are discussed.


Random Vector Canonical Form Stochastic System Internal Model Minimal Realization 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    ANDERSON B.D.O. "The inverse problem of stationary covariance generation" J. Stat. Phys. 1, 133–147, (1969).Google Scholar
  2. [2]
    COMREY A.L., A first Course in Factor Analysis Ac. Press, 1973.Google Scholar
  3. [3]
    ERSHOV M.P. "Extension of measures and stochastic equations" Theory Prob. Appl. XIX, 3, 431–444, (1974).Google Scholar
  4. [4]
    ERSHOV M.P. "Non anticipating solutions of stochastic equations" Proc. 3rd Japan-USSR Symp. on Probability Theory, Springer Lect. Notes Math 550, 655–691, (1976).Google Scholar
  5. [5]
    FAURRE P., CLERGET M., GERMAIN F., Opérateurs rationnels positifs, Dunod, 1979.Google Scholar
  6. [6]
    KALMAN R.E. "Identifiability and modeling in econometrics" in Developments in statistics, 4, P. Krishnaiah ed., 97–134, Ac Press, 1983.Google Scholar
  7. [7]
    KALMAN R.E. "Theory of modeling" Proc. IBM Syst. Science Symp., Oiso Japan, Y. Nishikawa ed. 53–69, (1979).Google Scholar
  8. [8]
    KENDALL M., STUART A., The Advanced Theory of Statistics, vol. III, Griffin 1976.Google Scholar
  9. [9]
    LINDQUIST A., PICCI G. "On the stochastic realization problem", SIAM J. Control and Optim. 17, 365–389, (1979).Google Scholar
  10. [10]
    LINDQUIST A., PICCI G. "Realization theory for multivariate stationary gaussian processes I: State space construction", Proc. 4th Intern. Symp. Math. Theory of Networks and Systems, July 1979, Delft, Holland, 140–148, (1979).Google Scholar
  11. [11]
    LINDQUIST A., PICCI G. "Realization theory for multivariate stationary gaussian processes II: State space theory revisited and dynamical representations of finite dimensional state spaces", Proc. 2nd Intern. Conf. on Information Sciences and Systems, Patras, Greece, July 1979, Reidel Publ., Co., 108–129, (1979).Google Scholar
  12. [12]
    LINDQUIST A., PICCI G. "State space models for gaussian stochastic processes", Stochastic Systems: The Mathematics of Filtering and Identification and Applications, M. Hazewinkel and J.C. Willems, Eds., Reidel Publ. Co., (1981).Google Scholar
  13. [13]
    LINDQUIST A., PICCI G., RUCKEBUSCH G. "On minimal splitting subspaces and Markovian representations", Math. Syst. Theory 12, 271–279, (1979).Google Scholar
  14. [14]
    MEHRA R. "Identification and estimation of the error-in-variables model (EVM) in structural form" Math. Prog. Study 5, 191–210, (1976).Google Scholar
  15. [15]
    NEVEU J., Processus Aléatoires Gaussiens, Presses de l'Université de Montreal, 1968.Google Scholar
  16. [16]
    PICCI G. "The stochastic realization problem" Proc. Symp. Sistemi Dinamici Stocastici, Rome June 1982, G. del Grosso ed. (1982).Google Scholar
  17. [17]
    RÜCKEBUSCH G. "Répresentations Markoviennes de processus gaussiens stationnaires" C.R. Acad. Sc. Paris, Serie A, 282, 649–651, (1976).Google Scholar
  18. [18]
    SÖDESTROM T. "Some methods for identification of linear systems with noisy input-output data", Proc. 5th IFAC Symp. Identif. Syst. Param. Estim., Darmstadt 1979, 1, 357–363, (1979).Google Scholar
  19. [19]
    Van PUTTEN C., Van SCHUPPEN J.H. "The weak and strong gaussian probabilistic realization problem" J. Multivar. Anal. 13, 118–137, (1983).Google Scholar
  20. [20]
    WILLEMS J.C. "From time series to linear systems" talk presented at the 6th Math. Th. of Networks and Systems Symp., Beer Sheva, Israel, June 1983.Google Scholar

Copyright information

© Springer-Verlag 1984

Authors and Affiliations

  • Lorenzo Finesso
    • 1
  • Giorgio Picci
    • 2
  1. 1.Ladseb-C.N.R.PadovaItaly
  2. 2.Istituto di Elettrotecnica ed ElettronicaPadovaItaly

Personalised recommendations