Regression operator in infinite dimensional vector spaces and its application to some identification problems
The main purpose of this paper was the transfer of the notion of regression into the theory of Banach valued random variables. This approach enables the formulation of identification problem in the domain of the infinite dimensional probability theory.
In the special case, when the random variables are Hilbert valued we state that the regression operator minimizes the square mean error. This makes the proof of the approximation theorem possible in the case of the linear bounded regression operator. Given further the statistical approximation theorem enables practically the identification of linear systems.
The problem unsolved in this paper is as well the approximation of the nonlinear regression operator.
It seems to the author that the general formulation of identification problem presented in his paper should help with solving also these problems.
KeywordsConditional Expectation Approximation Theorem Separable Hilbert Space Separable Banach Space Borelian Mapping
Unable to display preview. Download preview PDF.
- Dunford N., Schwartz J. — Linear Operators, Part I, London 1958Google Scholar
- Goldberger A.S. — Econometric Theory, John Wiley and Sons. Inc., New York-London-Sydney 1964Google Scholar
- Hille E. —Functional Analysis and Semi-groups, New York 1948Google Scholar
- Mourier E. — Les element aleatoires dans un espace de Banach, Annals de L'Institut Henri Poincare, 13(1953), pp. 161–244Google Scholar
- Rieffel N. — The Radon-Nikodym theorem for Bochner Integral, Transactions of the American Mathematical Society, 131 (1968), pp. 466–487Google Scholar
- Renyi A. —Probability theory, Academia Kiado, Budapest 1972Google Scholar