|
Abstract.
Monitoring the state of dynamic systems requires rapid data processing in real time, however, traditional methods such as the least squares method are not always efficient in the presence of outliers and contaminated data. The goal of this work is to implement and evaluate the computational efficiency of dynamic dynamic descent along nodal straight lines algorithms based on the least absolute deviations method for continuous monitoring tasks of dynamic regression models. We proposed dynamic coordinate-wise and gradient descent descent along nodal lines algorithms to implement the least absolute deviations method, the weighted and the generalised least absolute deviations methods. The dynamic implementation significantly improves performance compared to the static one, bringing the analysis time closer to that of the least squares method. Approbation on stock price data showed high estimation accuracy even with non-optimal initial data, but the descent along nodal straight lines showed minor deviations from the exact solution in the presence of multicollinearity, which are decreased when we used nonlinear models. Due to high performance and robustness to outliers, the algorithms can effectively solve the problems of continuous monitoring of rapidly changing processes.
Keywords:
linear regression, dynamic, robustness, least absolute deviations method, descent along nodal straight lines, monitoring.
DOI 10.14357/20718632250205
EDN LPFLZN
PP. 51-63.
References
1. Civil defense: Encyclopedia in 4 volumes. Vol. II (K – O). Moscow: FGBU VNII GOCHS (FTS); 2015. 624 p. (In Russ). 2. Lok P., Crowford J. The application of a diagnostic model and surveys in organizational development. Journal of Managerial Psychology. 2000;15(2):108–124. doi: 10.1108/02683940010310319. 3. Kolobov A.B. Vibrodiagnostics: Theory and practice. Moscow: Infra-Inzheneriia; 2019. 252 p. (In Russ.). 4. Shikhalev D.V. Problems of managing the fire safety system of a facility. Part II: monitoring methods. Control Sciences. 2022; No. 2:2–8. doi: 10.25728/cs.2022.2.1. 5. Nazarova I.B. Monitoring of the population health and health risk factors (Research methodology). RUDN Journal of Sociology. 2022;22(3):616–629 (In Russ). doi: 10.22363/2313-2272-2022-22-3-616-629. 6. Presnyakov V. Parameters and indicators of enterprise status monitoring. Economics and Mathematical Methods. 2022;58(3):70–78 (In Russ). doi: 10.31857/S042473880021700-9. 7. Akimov P. A., Matasov A.I. An iterative algorithm for l1-norm approximation in dynamic estimation problems. Automation and Remote Control. 2015;76(5):733–748. doi: 10.1134/S000511791505001X. 8. Lycheva M., Mironenkov A., Kurbatskii A., Fantazzini D. Forecasting oil prices with penalized regressions, variance risk premia and Google data. Applied Econometrics. 2022;4(68):28–49. doi: 10.22394/1993-7601-2022-68-28-49. 9. Domnina S.V., Savoskina E.V., Solopova N.A. Using regression models to analyse and forecast the residential property market. Fundamental research. 2024; No. 4:36–41 (In Russ). doi: 10.17513/fr.43591. 10. Hoffmann J.P. Linear regression models. Applications in R. CRC Press; 2022. 437 p. doi: 10.1201/9781003162230 11. Linnik Yu.V. Method of least squares and foundations of mathematical and statistical theory of observation processing: 2nd ed. Moscow: Fizmatgiz; 1962. 352 p. (In Russ). 12. Greene W.H. Econometric Analysis: 8th ed. Pearson; 2020. 1176 p. 13. Bhatia S., Frangioni J.V., Hoffman R.M., Iafrate A.J., Polyak K. The Challenges Posed by Cancer Heterogeneity. Nature Biotechnology. 2012;30(7):604–610. doi: 10.1038/nbt.2294. 14. Manevich V., Peresetsky A., Pogorelova P. Stock market and cryptocurrency market volatility. Applied Econometrics. 2022;65(1):65–76 (In Russian). doi: 10.22394/1993-7601-2022-65-65-76. 15. Ives A.R. Random errors are neither: On the interpretation of correlated data. Methods in Ecology and Evolution. 2022;13(10):2092–2105. doi: 10.1111/2041-210X.13971 16. Wan Ji-Zhong, Wang Chun-Jing, Marquet P.A. Environmental Heterogeneity as a Driver of Terrestrial Biodiversity on a Global Scale. Progress in Physical Geography: Earth and Environment. 2023; 47(6):912–930. doi: 10.1177/03091333231189045. 17. Mudrov V.I., Kushko V.L. Methods of measurement processing. Quasi-likelihood estimates. Moscow: Radio i sviaz'; 1983. 304 p. (In Russ). 18. Dodge Y. The concise encyclopedia of statistics. Springer; 2008. 616 p. 19. Boldin M.V., Simonova G.I., Tiurin Yu.N. Sign-based statistical analysis of linear models. Moscow: Nauka. Fizmatlit; 1997. 288 p. (In Russ). 20. Pan B., Chen M., Wang Y. Weighted least absolute deviations estimation for periodic ARMA models. Acta Mathematica Sinica, English Series. 2015;31:1273–1288. doi: 10.1007/s10114-015-4372-8. 21. Golovanov О.А., Tyrsin A.N. Linear regression analysis based on weighted least absolute deviation method: exact algorithms. Applied Mathematics and Control Sciences. 2024; No. 4: 52–64 (In Russ). doi: 10.15593/2499-9873/2024.4.04. 22. Tyrsin A.N., Sokolov L.A. The estimation of linear regression is based on the generalized least modules method. Journal of Samara State Technical University, Ser. Physical and Mathematical Sciences. 2010;5(21):134–142 (In Russ). 23. Armstrong R.D., Kung D.S. Algorithm AS132: Least absolute value estimates for a simple linear regression problem. Applied Statistics. 1978;27(3):363–366. 24. Tyrsin A.N. Algorithms for descent along nodal straight lines in the problem of estimating regression equations using the least absolute deviations method. Industrial Laboratory. 2021;87(5):68–75 (In Russ). doi: 10.26896/1028-6861-2021-87-5-68-75. 25. Panyukov A.V., Mezal Ya.A. Parametric identification of quasilinear difference equation. Bulletin of the South Ural State University, series Mathematics. Mechanics. Physics. 2019;11(4):32–38 (In Russ). doi: 10.14529/mmph190404 26. Panyukov A.V., Mezaal Ya.A. Stable identification of linear autoregressive model with exogenous variables on the basis of the generalized least absolute deviation method. Bulletin of the South Ural State University. Series: «Mathematical Modelling, Programming and Computer Software». 2018;11(1):35–43. doi: 10.14529/mmp180104. 27. Golovanov O.A., Tyrin A.N. Improving the speed of the generalized least absolute deviations method algorithm by refining the solution area. Proceedings of the international VSMS “Modern methods of the theory of boundary value problems. Pontryagin Readings – XXXIV”, Voronezh, 2023. 115–117 (In Russ). 28. Tyrsin A.N., Azaryan A.A. Exact evaluation of linear regression models by the least absolute deviations method based on the descent through the nodal straight lines. Bulletin of the South Ural State University, series «Mathematics. Mechanics. Physics». 2018;10(2):47–56 (In Russ). doi: 10.14529/mmph180205. 29. Azaryan A.A. Fast algorithms of modelling of multivariate linear regression dependences on the basis of the least absolute deviations method. Cand.Sc. Diss. Ekaterinburg. 2018. 148 p. (In Russ). 30. Golovanov O.A., Tyrsin A.N. Мodified gradient descent algorithm along nodal straight lines in regression analysis problem. Industrial Laboratory. Diagnostics of Materials. 2025; 91(3): 83–92 (In Russ). doi: 10.26896/1028-6861-2025-91-3-83-92. 31. Golovanov O.A., Tyrsin A.N. Regression analysis of data based on the method of least absolute deviations in dynamic estimation problems. Industrial Laboratory. Diagnostics of Materials. 2023;89(5):71–80 (In Russ). doi: 10.26896/1028-6861-2023-89-5-71-80. 32. Tukey J.W. A Survey of Sampling from Contaminated Distribution. Contributions to Probability and Statistics. Stanford: Stanford University Press; 1960. pp. 443-485. 33. Huber P.J., Ronchetti E.M. Robust Statistics: 2nd edition. Wiley; 2009. 363 p. 34. Clauset A., Shalizi C.R., Newman M.E.J. Power-law distributions in empirical data. SIAM review. 2009; 51(4): 661-703. doi: 10.1137/070710111. 35. FINAMP: Export of data of the joint-stock company Gazprom. Available from: https://www.finam.ru/quote/moex/gazp/export/?ysclid=m4v7f0tg2e936318213 [Accessed 20 December 2025].
|
