A.S. Babin1, M.I. Baryshnikov2, V.A. Galkin3, Yu.E. Gapanyuk4

1–4 Bauman Moscow State Technical University (Moscow, Russia)
 

The task of supervised learning continues to be the most significant task of machine learning. Mankind has accumulated a huge number of data sets, on the basis of which it is possible to build predictive models by solving problems of regression, classification, and time series forecasting. The practical solution of machine learning problems is impossible without the implementation of effective software libraries aimed at solving these problems. The group method of data handling is considered one of the most effective methods for solving the problem of supervised learning, while there are a few effective software libraries aimed at automating this method. The aim of the work is to increase the efficiency of solving the problem of supervised learning by implementing a software library aimed at automating the group method of data handling. The article analyzes in detail the general approach, which is the basis of the group method of data handling, as well as specific algorithms for implementing the method: COMBI, MULTI, MIA, RIA. The creation of an effective software library aimed at implementing the COMBI, MULTI, MA, RIA algorithms is considered. Experiments have been carried out for regression and time series forecasting problems, showing the advantages of the developed library. Using the developed library, it is possible to effectively solve the problems of regression, classification, forecasting of time series by the group method of data handling.

Babin A.S., Baryshnikov M.I., Galkin V.A., Gapanyuk Yu.E. Development of a library for the group method of data handling. Dynamics of complex systems. 2023. V. 17. № 2. P. 51−65. DOI: 10.18127/j19997493-202302-06 (in Russian)

1. Zhuk D.M., Volosatova T.M., Spasenov A.Yu., Kucherov K.V. Ocenka dinamicheskih sistem s ispol'zovaniem modal'no-lingvisticheskogo analiza mnogomernyh vremennyh ryadov. Dinamika slozhnyh sistem – XXI vek. 2020. T. 14. № 1. S. 39–45. DOI 10.18127/j19997493-202001-04 (in Russian).
2. Myshenkov K.S., Nekula H. Ispol'zovanie metodov mashinnogo obucheniya dlya prognozirovaniya nevrologicheskih zabolevanij. Dinamika slozhnyh sistem – XXI vek. 2022. T. 16. № 1. S. 66–74. DOI 10.18127/j19997493-202201-07 (in Russian).
3. Ivahnenko A.G. Sistemy evristicheskoj samoorganizacii v tekhnicheskoj kibernetike. Kiev: Tekhnika. 1971. 372 s. (in Russian).
4. Ivakhnenko A.G. Polynomial Theory of Complex Systems. Reprinted by permission from IEEE Transactions on Systems, Man, and Cybernetics. October 1971. V. SMC-1. № 4. P. 364–378.
5. Schmidhuber J. Deep Learning in Neural Networks: An Overview. Neural Networks. 2015. V. 61. P. 85–117.
6. Dag O., Ceylan Y. GMDH: An R Package for Short Term Forecasting via GMDH-Type Neural Network Algorithms. The R Journal. Aug. 2016. V. 8/1. P. 379–386.
7. GmdhPy library. URL: https://github.com/kvoyager/GmdhPy (data obrashcheniya: 13.03.2023).
8. Ponyatskij V.M., Veleshki S.I., Zhirnova A.V. Ispol'zovanie metoda gruppovogo ucheta argumentov dlya vybora struktury modeli dinamicheskogo ob"ekta. Izv. Tul'skogo gosudarstvennogo universiteta. Tekhnicheskie nauki. 2013. № 2. S. 255–267 (in Russian).
9. Stepashko V.S., Kostenko Yu.V. Issledovanie svojstv kombinatorno-selekcionnogo (mnogoetapnogo) algoritma MGUA. Modelirovanie i upravlenie sostoyaniem ekologo-ekonomicheskih sistem regiona: sb. nauch. tr. Kiev: IK NANU. 2001. S. 69–76 (in Russian).
10. Yasinskij I.F. i dr. O sozdanii nejrosetevoj gibridnoj sistemy dlya professional'nogo orientirovaniya studentov. Vestnik Cherepoveckogo gosudarstvennogo universiteta. 2021. № 5 (104). S. 59–71 (in Russian).
11. Ivahnenko A.G., Stepashko V.S. Pomekhoustojchivost' modelirovaniya. Kiev: Naukova dumka. 1985. 216 s. (in Russian).
12. Pidnebesna H., Savchenko-Synyakova Y., Stepashko V. Application of the Ontological Approach to the Iterative GMDH Algorithms Metamodel Construction. IEEE 17th International Conference on Computer Sciences and Information Technologies (CSIT). 2022. P. 576–579.
13. Boost C++ libraries. URL: https://www.boost.org (дата обращения: 13.03.2023).
14. Eigen library. URL: https://github.com/PX4/eigen (дата обращения: 13.03.2023).
15. Indicators library. URL: https://github.com/p-ranav/indicators (дата обращения: 13.03.2023).
16. Pybind11 library. URL: https://github.com/pybind/pybind11 (дата обращения: 13.03.2023).
17. Eigen library: Linear algebra and decompositions. URL: https://eigen.tuxfamily.org/dox/group__TutorialLinearAlgebra.html (дата обращения: 13.03.2023).
18. Laptop Price Dataset. URL: https://www.kaggle.com/datasets/muhammetvarl/laptop-price (дата обращения: 13.03.2023).
19. Weather Madrid 1997-2015 Dataset. URL: https://www.kaggle.com/ datasets/juliansimon/weather_madrid_lemd_1997_2015.csv (дата обращения: 13.03.2023).