Skip to main content
Log in

Building correct estimation algorithms as a constrained optimization problem

  • Published:
Computational Mathematics and Mathematical Physics Aims and scope Submit manuscript

Abstract

For estimation algorithms, the problem of building correct algorithms by modifying weights of features and weights of objects is examined. Criteria for the possibility to build a correct algorithm are obtained for certain cases. Conditions of the possibility to build a correct algorithm are obtained in terms of solving a constrained optimization problem. An optimization method is proposed. Under these conditions, the proposed method significantly reduces the computational complexity of synthesizing a correct algorithm.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

References

  1. Yu. I. Zhuravlev, “Correct Algorithms over Sets of Incorrect (Heuristic) Algorithms: Part II,” Kibernetika, No. 6, 21–27 (1977).

  2. Yu. I. Zhuravlev, “Correct Algorithms over Sets of Incorrect (Heuristic) Algorithms: Part III,” Kibernetika, No. 2, 35–43 (1978).

  3. Yu. I. Zhuravlev and I. V. Isaev, “Constructing Recognition Algorithms That are Correct for a Given Validation Sample” Zh. Vychisl. Mat. Mat. Fiz. 19, 726–738 (1979).

    MATH  MathSciNet  Google Scholar 

  4. K. V. Rudakov, On Certain Classes of Recognition Algorithms (General Results) (Vychisl. Tsentr Akad. Nauk SSSR, Moscow, 1980) [in Russian].

    Google Scholar 

  5. K. V. Rudakov, On Certain Classes of Recognition Algorithms (Parametric Models) (Vychisl. Tsentr Akad. Nauk SSSR, Moscow, 1981) [in Russian].

    Google Scholar 

  6. V. V. Ryazanov, “Optimization of Estimation Algorithms Using Parameters Describing the Representative Property of Reference Rows,” Zh. Vychisl. Mat. Mat. Fiz. 16, 1559–1570 (1976).

    MATH  Google Scholar 

  7. V. V. Ryazanov, “Optimal Group Decisions in Recognition and Classification,” Doctoral Dissertation in Mathematics and Physics (VTs RAN, Moscow, 1994).

    Google Scholar 

  8. N. N. Katerinochkina, “Methods for Finding a Maximal Consistent Subsystem of a System of Linear Inequalities,” in Communications in Applied Mathematics (Vychisl. Tsentr, Ross. Akad. Nauk, Moscow, 1997) [in Russian].

    Google Scholar 

  9. B. T. Polyak, Introduction to Optimization (Nauka, Moscow, 1983) [in Russian].

    MATH  Google Scholar 

  10. S. B. Pshenichnikov and A. G. Shtern, “Methods for Solving Systems of Nonlinear Algebraic Equations,” in Mathematical Methods and Computer-Aided Systems in Geology: A Survey (Moscow, 1995) [in Russian].

  11. S. B. Pshenichnikov, “Spinor Elimination and Formulae for Relations between Unknowns in Systems of Nonlinear Algebraic Equations,” Latv. Mat. Ezhegodnik, No. 30, 150–161 (1986).

  12. G. A. Zaitsev, Algebraic Problems in Mathematical and Theoretical Physics (Nauka, Moscow, 1974) [in Russian].

    Google Scholar 

  13. K. A. Rybnikov, Introduction to Combinatorial Analysis (Mosk. Gos. Univ., Moscow, 1985) [in Russian].

    MATH  Google Scholar 

  14. J. Edmonds, “Matroids and the Greedy Algorithm,” Math. Program. 1, 127–136 (1971).

    Article  MATH  MathSciNet  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to A. Yu. Smetanin.

Additional information

Original Russian Text © A.Yu. Smetanin, 2008, published in Zhurnal Vychislitel’noi Matematiki i Matematicheskoi Fiziki, 2008, Vol. 48, No. 3, pp. 529–535.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Smetanin, A.Y. Building correct estimation algorithms as a constrained optimization problem. Comput. Math. and Math. Phys. 48, 500–506 (2008). https://doi.org/10.1134/S0965542508030123

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1134/S0965542508030123

Keywords

Navigation