ACCOUNTING FOR A GIVEN ERROR LEVEL IN ESTIMATION OF PARAMETERS OF A PIECEWISELINEAR REGRESSION MODEL
- Authors: Noskov S.I.1, Belyaev S.V.1, Bychkov Y.A.2
-
Affiliations:
- Irkutsk State Transport University
- Иркутский государственный университет путей сообщения
- Issue: No 2 (2025)
- Pages: 75-84
- Section: MODELS, SYSTEMS, MECHANISMS IN THE TECHNIQUE
- URL: https://journal-vniispk.ru/2227-8486/article/view/307554
- DOI: https://doi.org/10.21685/2227-8486-2025-2-6
- ID: 307554
Cite item
Full Text
Abstract
Background. The development of mathematical models of complex objects is usually accompanied by an analysis of their admissibility using both strict formal criteria and procedures, and various heuristic techniques. This applies to models of any type, including regression. The aim of the study is to develop an algorithmic method for identifying parameters of the Leontiev piecewise linear regression model with the maximum number of admissible approximation errors. This number can be one of the criteria for assessing the adequacy (admissibility) of regression models. Materials and methods. To achieve the stated goal, the mathematical apparatus for solving linear Boolean programming problems was used. Results. The formulated problem is reduced to a linear Boolean programming problem of a dimension acceptable for real objects. Conclusions. The approach described in the work allows for an acceptable level of admissibility of approximation errors in a piecewise linear regression model. An adequate regression model of the aluminum industry of the Russian Federation has been constructed.
About the authors
Sergei I. Noskov
Irkutsk State Transport University
Email: sergey.noskov.57@mail.ru
Doctor of technical sciences, professor, professor of the sub-department of information systems and information security
(15 Chernyshevskogo street, Irkutsk, Russia)Sergey V. Belyaev
Irkutsk State Transport University
Email: bsv2001@list.ru
Master degree student of the sub-department of information systems and information security
(15 Chernyshevskogo street, Irkutsk, Russia)Yuriy A. Bychkov
Иркутский государственный университет путей сообщения
Author for correspondence.
Email: nik24-11@mail.ru
аспирант кафедры информационных систем и защиты информации
(Россия, г. Иркутск, ул. Чернышевского, 15)References
- Xu R.-Y., Zhang P., Dai S.-Q., Wong S.C. Admissibility of a Wide Cluster Solution in "Anisotropic" Higher-Order Traffic Flow Models. SIAM Journal on Applied Mathematics. 2007;68:562–573.
- Yi˘git A., Tun C. On the stability and admissibility of a singular differential system with constant delay. International Journal of Mathematics and Computer Science. 2020;15:641–660.
- Zhuang G., Xia J., Sun W. et al. Asynchronous admissibility and fault detection for delayed implicit Markovian switching systems under hidden Markovian model mechanism. International Journal of Robust and Nonlinear Control. 2021;31:7261–7279.
- Zhen B., Weber K., Mejia-Ramos J.P. Mathematics Majors’ Perceptions of the Admissibility of Graphical Inferences in Proofs. International Journal of Research in Undergraduate Mathematics Education. 2016;2:1–29.
- Berger J.O., Strawderman W.E. Choice of Hierarchical Priors: Admissibility in Estimation of Normal Means. The Annals of Statistics. 1996;24:931–951.
- Jiao T., Zong G., Pang G. et al. Admissibility analysis of stochastic singular systems with Poisson switching. Applied Mathematics and Computation. 2020;386:1–14.
- Biagini S., Černý A. Admissible Strategies in Semimartingale Portfolio Selection. SIAM Journal on Control and Optimization. 2011;49:42–72.
- Benamar M.N., Ghezzar M.A., Bouagada D., Benyettou K. On the admissibility and robust stabilization of 2D singular continuous–discrete linear systems. International Journal of Dynamics and Control. 2024;12:1728–1742.
- Leeb H., Kabaila P. Admissibility of the Usual Confidence Set for the Mean of a Univariate or Bivariate Normal Population: The Unknown Variance Case. Journal of the Royal Statistical Society. Series B: Statistical Methodology. 2017;79:801–813.
- Noskov S.I. Tekhnologiya modelirovaniya ob"ektov s nestabil'nym funktsionirovaniem i neopredelennost'yu v dannykh = Technology for modeling objects with unstable functioning and uncertainty in data. Irkutsk: Oblinformpechat', 1996:320. (In Russ.)
- Noskov S.I. Implementation of a competition of regression models using the criterion of consistency of behavior. Vestnik Voronezhskogo gosudarstvennogo universiteta. Seriya: Sistemnyy analiz i informatsionnye tekhnologii = Bulletin of Voronezh State University. Series: System Analysis and Information Technology. 2021;(2):153–160. (In Russ.)
- Pyakillya B.I., Zhmud' V.A. Correct identification of an object with a delay for its management. Avtomatika i programmnaya inzheneriya = Automation and software engineering. 2015;(3):51–57. (In Russ.)
- Smetanin Yu.G., Ul'yanov M.V. An algebraic structure with partial operations and a computational model for arithmetic of bounded non-negative integers. Vychislitel'nye tekhnologii = Computing technologies. 2013;18(4):48–63. (In Russ.)
- Kaganovich B.M., Stennikov V.A. Development of the principles and models of classical equilibrium thermodynamics and their applications in energy research. Izvestiya Rossiyskoy akademii nauk. Energetika = Proceedings of the Russian Academy of Sciences. Energy industry. 2018;(6):76–87. (In Russ.)
- Rybakov V.V. Dynamic time operations in multi-agent logics. Algebra i logika = Algebra and Logic. 2022;61(5):600–618. (In Russ.)
- Efremov A.A. Generalized projection operator for solving problems of program motion stabilization. Izvestiya Tul'skogo gosudarstvennogo universiteta. Tekhnicheskie nauki = Proceedings of Tula State University. Technical sciences. 2024;(7):230–235. (In Russ.)
- Noskov S.I., Lonshakov R.V. Identification of piecewise linear regression parameters. Informatsionnye tekhnologii i problemy matematicheskogo modeli-rovaniya slozhnykh system = Information technologies and problems of mathematical modeling of complex systems. 2008;(6):63–64. (In Russ.)
- Noskov S.I., Khonyakov A.A. Piecewise linear regression models of passenger transportation volumes by long-distance transport. Modeli, sistemy, seti v ekonomike, tekhnike, prirode i obshchestve = Models, systems, and networks in economics, technology, nature, and society. 2021;(4):80–89. (In Russ.)
- Noskov S.I., Zhukova M.S., Kirillova T.K. et al. Refinement of methods for identifying parameters of some piecewise linear regressions. Nauchnye trudy KubGTU = Scientific works of KubSTU. 2023;(2):75–81. (In Russ.)
- Noskov S.I., Shakhurov A.N. Maximizing the number of acceptable approximation errors when building a linear regression model. Vestnik Yugorskogo gosudarstvennogo universiteta = Bulletin of Ugra State University. 2024;20(3):57–62. (In Russ.)
- Rossiyskiy statisticheskiy ezhegodnik. 2010: stat. sb. = Russian Statistical Yearbook. 2010 : stat. collecting. Moscow: Rosstat, 2010:813. (In Russ.)
- Rossiyskiy statisticheskiy ezhegodnik. 2016: stat. sb. = Russian Statistical Yearbook. 2016 : stat. collecting. Moscow: Rosstat, 2016:725. (In Russ.)
- Rossiyskiy statisticheskiy ezhegodnik. 2018: stat. sb. = Russian Statistical Yearbook. 2018 : stat. collecting. Moscow: Rosstat, 2018:694. (In Russ.)
- Rossiyskiy statisticheskiy ezhegodnik. 2022: stat. sb. = Russian Statistical Yearbook. 2022 : stat. collecting. Moscow: Rosstat, 2022:691. (In Russ.)
- Rossiyskiy statisticheskiy ezhegodnik. 2023: stat. sb. = Russian Statistical Yearbook. 2023 : stat. collecting. Moscow: Rosstat, 2023:701. (In Russ.)
Supplementary files
