Development of a Regularized Bayesian Toolkit for the System Management of Small and Medium-sized Enterprises and its Testing Based on the Construction of an Adaptive Model
( Pp. 33-46)

The main task in building complex predictive models is to prevent overfitting, especially in small samples, which can lead to incorrect management conclusions. This is due to the high dynamics of the market, limited resources and the nature of information flows (incompleteness, noise) in small and medium-sized enterprises (SMEs), which require the creation of stable and accurate intelligent tools to support management decisions. To solve this problem, data mining uses regularization extensively, which, in fact, imposes restrictions on the complexity of the model. The combination of Bayesian inference and regularization principles forms a regularizing Bayesian approach that creates stable and generalizing models. The role of Bayesian regularization is to select weighting factors for forecasting accuracy, and when generalized, this approach becomes the basis for creating digital platforms for managing complex systems in Industry 4.0 and high-tech SMEs. Thus, the relevance of the research in this article is due to the need to develop an integrated methodological framework that combines the principles of system analysis with a regularizing Bayesian approach. The research aims to contribute to improving the quality of management decisions in SMEs through the introduction of intelligent, statistically sound, and data incompleteness–resistant models that have been integrated into the new adaptive hierarchical regularizing Bayesian model of decision-making. AIRBM successfully implements the principle of system analysis, linking external system uncertainty with the internal configuration of the model, which provides a reliable and stable basis for intelligent management of SMEs in the digital economy.

Zvyagin L.S. Development of a regularized Bayesian toolkit for the system management of small and medium-sized enterprises and its testing based on the construction of an adaptive model. Computational Nanotechnology. 13, 1 (2026), 33–46 DOI: 10.33693/2313-223X-2026-13-1-33-46. EDN: MCLOCR

Bormotov I.V., Zhukov R.A., Prokopchina S.V. Methodology of Bayesian intelligent measurements for information systems and its application to the study of the value world of Russian youth. In: Proceedings of the XXVI International Conference on Soft Computing and Measurements (SCM). In 2 vols. St. Petersburg: ETU “LETI”, 2023. Vol. 1. Pp. 3–6.
Dmitryuk T.G., Zori S.A. System analysis of functional features of an enterprise as a management object. Bulletin of Donetsk National University. Series G: Technical Sciences. 2024. No. 2. (In Rus.). Pp. 137–153. DOI: 10.5281/zenodo.14514826.
Evseeva O.A. Managing the development of small and medium-sized industrial enterprises using system analysis methods. Ivestia Vuzov. Series: Economics, Finance and Production Management. 2015. No. 1 (23). Pp. 75–80. (In Rus.)
Zaytsev S.V. Applied system analysis and systems engineering as tools for optimizing economic processes. Soft Measurements and Computing. 2022. Vol. 52. No. 3. Pp. 70–84. (In Rus.). DOI: 10.36871/2618-9976.2022.03.008.
Zaslavskaya V.L., Zaslavskiy R.K., Prokopchina S.V. Intelligent processing of big data in small business tasks based on Bayesian intelligent technologies. Soft Measurements and Computing. 2022. Vol. 61. No. 12. Pp. 65–74. (In Rus.). DOI: 10.36871/2618-9976.2022.12.005.
Zolotarevskiy A.Yu. Analytics and decision support system for a small manufacturing enterprise based on the regularizing Bayesian approach. Soft Measurements and Computing. 2024. Vol. 75. No. 2-2. Pp. 38–48 (In Rus.).. DOI: 10.36871/2618-9976.2024.02-2.004.
Kozhomberdieva G.I., Burakov D.P., Khamchichev G.A. Development of programs for decision support based on Bayesian probabilistic models. Software Products and Systems. 2022. Vol. 35. No. 2. Pp. 184–194. (In Rus.)
Prokopchina S.V. Intelligent sensor networks in Industry 5.0. A generalized concept for creating digital platforms for managing complex systems based on a regularizing Bayesian approach. In: Proceedings of the XXV International Conference on Soft Computing and Measurements (SCM). St. Petersburg: ETU “LETI”, 2022. Vol. 1. Pp. 3–10.
Sharabura E.A., Tin Yu.A. Application of system analysis in solving problems of small business. In: Modern problems of economic development of enterprises, industries, complexes, territories. Proceedings of the International Scientific and Practical Conference. A.E. Zubarev et al. (eds.) Khabarovsk: Pacific State University, 2020. Pp. 429–432.
Abonguie D.F., Nyam Y.S., Hoeyi P.K. A systematic analysis of systems thinking and the sustainability of small and medium size enterprises (SMEs): A global evaluation. Scientific African. 2025. Vol. 28. Art. e02738. DOI: 10.1016/j.sciaf.2025.e02738.
Hahn E., Doh J. Using Bayesian methods in strategy research: An extension of Hansen et al. Strategic Management Journal. 2006. Vol. 27. No. 8. Pp. 783–798. DOI: 10.1002/smj.539.
Liao R., Chai Y. Research on the business performance evaluation method for small and medium-sized enterprises in cross-border e-commerce based on artificial bee colony optimized LSTM model. Scientific Reports. 2025. Vol. 15. Art. 31698. DOI: 10.1038/s41598-025-17435-x.
Nuzhny A. Bayes regularization in the selection of weight coefficients in the predictor ensembles. Proceedings of the Institute for System Programming of the RAS. 2019. Vol. 31. No. 4. Pp. 113–120. DOI: 10.15514/ISPRAS-2019-31-4-7.
Wüthrich M.V., Merz M. Bayesian methods, regularization and expectation-maximization. In: Statistical foundations of actuarial learning and its applications. Cham: Springer, 2023. Pp. 393–440. DOI: 10.1007/978-3-031-12409-9_6.

systems analysis, small and medium-sized enterprises (SMEs), Bayesian approach, regularization, adaptive model, risk management.