OPTIMIZATION OF FACTOR VARIABLE SELECTION IN ECONOMETRIC MODELING: PRACTICAL RECOMMENDATIONS
DOI:
https://doi.org/10.31110/fmo2025.v40i5-04Keywords:
econometric modeling, factor variables, factor selection, multicollinearity, correlation analysisAbstract
Formulation of the problem. The correct selection of factor variables in modern econometric studies is a key condition for ensuring the accuracy, stability, and predictive power of models. However, student papers often contain typical errors related to an excessive number of factors, weak theoretical justification for their selection, or the presence of multicollinearity. The article substantiates the need to create practical recommendations that combines scientific validity with ease of use and ensures the correct selection of factor variables in the process of modeling economic phenomena.
Materials and methods. The study uses a set of theoretical, statistical, and econometric methods, among which correlation and regression analysis played a key role, allowing us to identify and quantitatively assess the relationships between variables. To increase the reliability of the results and the stability of the model, a multicollinearity check was performed using the Farrar–Glauber algorithm, which involves the use of χ², F, and t criteria to assess the interdependencies between independent variables.
Results. As a result of the study, a step-by-step algorithm for selecting factor variables is proposed, which involves: (1) calculating correlation coefficients to form a vector and correlation matrix; (2) selecting variables that have a strong relationship with the outcome variable and weak mutual correlation; (3) verification of the informativeness of variables using the method of information capacity indicators and the multiple correlation coefficient; (4) additional analysis for multicollinearity to identify redundant factors. The algorithm ensures the validity of the predictor choice, enhances the accuracy and stability of models, and is convenient for educational purposes.
Conclusion. The proposed approach enables the systematic organization of the process of selecting factor variables in student econometric research, combining clarity, accessibility, and scientific rigor. Its application contributes to the formation of students' analytical thinking skills, statistical interpretation of results, and a conscious approach to model building.
Downloads
References
Abramov, V., Astafieva, M., Boiko, M., Bodnenko, D., Bushma, A., Vember, V., Hlushak, O., Zhyltsov, O., Ilich, L., Kobets, N., Kovaliuk, T., Kuchakovska, H., Lytvyn, O., Lytvyn, P., Mashkina, I., Morze, N., Nosenko, T., Proshkin, V., Radchenko, S., & Yaskevych, V. (2021). Theoretical and practical aspects of the use of mathematical methods and information technology in education and science. https://doi.org/10.28925/9720213284km
Dudáš, A., Michalíková, A., & Jašek, R. (2025). Fuzzy masks for correlation matrix pruning. IEEE Access, 13, 35387–35400. https://doi.org/10.1109/access.2025.3544027
Desboulets, L. (2018). A review on variable selection in regression analysis. Econometrics, 6(4). https://doi.org/10.3390/econometrics6040045
Fan, J., Ke, Y., & Wang, K. (2018). Factor-adjusted regularized model selection. Econometrics: Econometric Model Construction. https://doi.org/10.2139/ssrn.3248047
García-Donato, G., & Paulo, R. (2021). Variable selection in the presence of factors: A model selection perspective. Journal of the American Statistical Association, 117(540), 1847–1857. https://doi.org/10.1080/01621459.2021.1889565
Hall, P., & Miller, H. (2009). Using generalized correlation to effect variable selection in very high dimensional problems. Journal of Computational and Graphical Statistics, 18(3), 533–550. https://doi.org/10.1198/jcgs.2009.08041
Kuo, L., & Mallick, B. (1998). Variable selection for regression models. Sankhyā: The Indian Journal of Statistics, Series B, 60(1), 65–81.
Kvalheim, O. (2020). Variable importance: Comparison of selectivity ratio and significance multivariate correlation for interpretation of latent-variable regression models. Journal of Chemometrics, 34(9). https://doi.org/10.1002/cem.3211
Obushnyi, S., et al. (2021). Ensuring data security in the peer-to-peer economic system of the DAO. In Cybersecurity Providing in Information and Telecommunication Systems II (Vol. 3187, pp. 284–292).
Obushnyi, S., et al. (2022). Autonomy of economic agents in peer-to-peer systems. In Cybersecurity Providing in Information and Telecommunication Systems (Vol. 3288, pp. 125–133).
Turlach, B., Venables, W., & Wright, S. (2005). Simultaneous variable selection. Technometrics, 47(3), 349–363. https://doi.org/10.1198/004017005000000139
Virovets, D., et al. (2023). Ways of interaction of autonomous economic agents in decentralized autonomous organizations. In Cybersecurity Providing in Information and Telecommunication Systems (Vol. 3421, pp. 182–190).
Hlushak, O. M., & Semenyaka, S. О. (2017). Economic and mathematical modeling is a promising area of applied mathematics. Physical and Mathematical Education, 1(11), 28–31. http://fmo-journal.fizmatsspu.sumy.ua/journals/2017-v1-11/2017_1-11-GlushakSemenyaka_Scientific_journal_FMO.pdf (in Ukrainian).
Hlushak, O. M, & Semenyaka, S. O. (2018). Prerequisites for building a multifactor econometric model: a study on multicollinearity. Physical and Mathematical Education, 1(15), 171-175. https://doi.org/10.31110/2413-1571-2018-015-1-031 (in Ukrainian).
Hlushak, O. M, Бодненко, Д., & Semenyaka, S. O. (2018). Developing the IT skills of future financiers while studying the discipline “Econometrics.” Educational discourse, 1–2(20–21), 325–340. Borys Grinchenko Kyiv University. http://od.kubg.edu.ua/index.php/journal/article/view/492 (in Ukrainian).
Downloads
Published
Issue
Section
Categories
How to Cite
License
Copyright (c) 2025 Оксана Глушак, Світлана Семеняка, Надія Зінченко
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
- Authors grant the journal a right of the first publication of the work under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (CC BY-NC-SA 4.0)that allows others freely to use (read, copy and print) submissions, search content and link to published articles, disseminate their full text and use them for any legitimate non-commercial purposes (i.e. educational or scientific) with the mandatory reference to the article’s authors and initial publication in this journal.
- Original published articles cannot be used by users (exept authors) for commercial purposes or distributed by third-party intermediary organizations for a fee.
