G.N. Kamyshova1, D.A. Prokopeva2

1,2 Financial University under the Government of the Russian Federation (Moscow, Russia)

1 gnkamyshova@fa.ru, 2 202090@edu.fa.ru

Due to the ever-increasing demand of the world population for food, food exports are acquiring strategic importance for the country's economy. Food exports depend on many parameters, so forecasting them has its own characteristics. Models based on artificial neural networks are finding increasing applications in solving such problems, so the development of approaches to the use of neural network models for forecasting agricultural exports, which allow taking into account various parameters, is a very urgent task. The purpose of the work is to propose a modern approach to forecasting based on artificial neural networks for regional agricultural exports. Neural network models for forecasting agricultural exports of Russian regions are proposed using the example of the Krasnodar Territory. The algorithm based on LSTM neural networks makes it possible to predict the volume of exports of agricultural products based on available statistical information and has a number of advantages compared to classical forecasting methods. Development of applied neural network models for forecasting regional agricultural exports, the advantages of which are the ability to take into account a large number of factors, adaptability and versatility. The study can be a useful tool for government agencies and business structures in making decisions on the directions of development of agricultural exports and budget planning.

Kamyshova G.N., Prokopeva D.A. Neural network models for forecasting regional agricultural exports. Neurocomputers. 2024. V. 26. № 6. Р. 6-13. DOI: https://doi.org/10.18127/j19998554-202406-02 (In Russian)

1. Statistics | FAO | Food and Agriculture Organization of the United Nations. [Electronic resource] – Access mode: https://www.fao.org/ statistics/ru/, date of reference 20.03.2024. (in Russian)
2. Agricultural exports of Russian regions in 2022 – Agroexport. [Electronic resource] – Access mode: https://aemcx.ru/2023/04/12/ag-rarnyj-eksport-regionov-rossii-v-2022-g/, date of reference 20.03.2024. (in Russian)
3. Magomadov E.M. Statistical analysis of trends in the development of Russia’s foreign trade. Economic and legal issues. 2022. № 167. P. 81–85. DOI 10.14451/2.167.81. (in Russian)
4. Kamyshova G.N. Statistical analysis of world grain trade. Scientific almanac. 2024. № 2-2(112). P. 91–94. (in Russian)
5. Kiselev S.V., Romashkin R.A., Belugin A.Yu. Russia’s agrifood exports until 2030: Projection from a partial equilibrium model. Journal of the New Economic Association 2022. № 4(56) P. 69–90. DOI 10.31737/2221-2264-2022-56-4-4. (in Russian)
6. Che N.X. Intelligent Export Diversification: An Export Recommendation System with Machine Learning. IMF Working Papers. 2020. V. 2020. № 175. P. 46. DOI 10.5089/9781513555959.001.
7. Kamyshova G.N., Zholobova G.N. An analysis of foreign trade in grains and machine learning methods in decision support for agrifood exports. International Research Journal. 2024. № 4(142) DOI 10.23670/IRJ.2024.142.9. (in Russian)
8. Federal Customs Service. [Electronic resource] – Access mode: https://customs.gov.ru/statistic, date of reference 10.02.2024. (in Russian)
9. Exports and imports of Russia by goods and countries. [Electronic resource] – Access mode: https://ru-stat.su/, date of reference 20.02.2024.
10. Sohrabpour V., Oghazi P., Toorajipour R., Nazarpour A. Export sales forecasting using artificial intelligence. Technological Forecasting and Social Change. 2021. V.163. P. 120480. DOI 10.1016/j.techfore.2020.120480.
11. Benderskaya E.N., Nikitin K.V. Features of computer modeling of impulse recurrent neural networks. Neurocomputers: development, application. 2014. № 2. P. 54–65. (in Russian)
12. Zhang G., Patuwo B.E., Hu M.Y. Forecasting with artificial neural networks: The state of the art. International Journal of Forecasting. 1998. V. 14. № 1. 35–62. DOI 10.1016/S0169-2070(97)00044-7.
13. Olah C. LSTM – long-term short-term memory networks. [Electronic resource] – Access mode: https://habr.com/ru/companies/wun-derfund/articles/331310 /, date of reference 20.02.2024.
14. Sezer O.B., Gudelek U., Ozbayoglu M. Financial Time Series Forecasting with Deep Learning: A Systematic Literature Review: 2005–2019. Applied Soft Computing. 2020. V. 90. P. 106181. DOI 10.1016/j.asoc.2020.106181.