Advertisement

The Algorithms of Extremal Parameter Grouping

  • Gintautas Dzemyda
Conference paper
Part of the Advances in Simulation book series (ADVS.SIMULATION, volume 1)

Abstract

The structural methods for the empirical data processing are used widely in systems analysis. The method of extremal parameter grouping [1,2] belongs to this class of methods. It is devoted to the partition of the parameters x1,…, xn into a fixed number p of the unintersec-ting groups A1,…,Ap. The correlation matrix \({\rm{R = }}\left\{ {{{\rm{r}}_{{{\rm{x}}_{\rm{i}}}{{\rm{x}}_{\rm{j}}}}}{\rm{,i,j = }}\overline {{\rm{l,n}}} } \right\}\) characterizes the connections among the parameters \({{\rm{(}}{{\rm{r}}_{{{\rm{x}}_{\rm{i}}}{{\rm{x}}_{\rm{j}}}}}}\) is the corre-lation coefficient of parameters xi. and xj.). The covariance matrix may be used instead of the matrix R. However, the parameters with a greater dispersion will have greater significance in the analysis.

Keywords

Function Depen Maximal Eigenvalue Structural Method Great Dispersion Initial Partition 
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.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    Braverman, E.M. and I.B. Muchnik: The structural methods for the empiric data processing. Nauka, Moscow 1983.Google Scholar
  2. [2]
    Dzemyda, G.: On the extremal parameter grouping. Teorija Optimal-jnych Reshenij, 12, Vilnius (1987), 28–42. (in Russian)MathSciNetMATHGoogle Scholar
  3. [3]
    Dzemyda, G.: LP-search with extremal problem structure analysis. Stochastic Control: Proceedings of the 2nd IPAC Symposium, IFAC Proceedings Series, 1987, Number 2, Pergamon Press, New York, 1987, 499–502.Google Scholar
  4. [4]
    Dzemyda, G. and J. Valevičiene: The extremal parameter grouping in cluster analysis. Teorija Optimaljnych Reshenij, 13, Vilnius (1988). (to appear in Russian)Google Scholar

Copyright information

© Akademie-Verlag Berlin 1988

Authors and Affiliations

  • Gintautas Dzemyda
    • 1
  1. 1.Institute of Mathematics and Cybernetics Lithuanian SSR Academy of SciencesVilniusUSSR

Personalised recommendations