<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.3 20210610//EN" "JATS-journalpublishing1-3.dtd">
<article article-type="research-article" dtd-version="1.3" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xml:lang="ru"><front><journal-meta><journal-id journal-id-type="publisher-id">dt</journal-id><journal-title-group><journal-title xml:lang="ru">Цифровая трансформация</journal-title><trans-title-group xml:lang="en"><trans-title>Digital Transformation</trans-title></trans-title-group></journal-title-group><issn pub-type="ppub">2522-9613</issn><issn pub-type="epub">2524-2822</issn><publisher><publisher-name>Educational Establishment “Belarusian State University of Informatics and Radioelectronics”</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="doi">10.35596/1729-7648-2026-32-2-44-52</article-id><article-id custom-type="elpub" pub-id-type="custom">dt-1039</article-id><article-categories><subj-group subj-group-type="heading"><subject>Research Article</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="ru"><subject>ТЕХНИЧЕСКИЕ НАУКИ</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="en"><subject>TECHNICAL SCIENCES</subject></subj-group></article-categories><title-group><article-title>Исследование влияния гиперпараметров на точность нейросетевого предсказания с использованием набора данных Fashion-MNIST</article-title><trans-title-group xml:lang="en"><trans-title>Investigation of the Impact of Hyperparameters on the Accuracy of Neural Network Predictions Using the Fashion-MNIST Dataset</trans-title></trans-title-group></title-group><contrib-group><contrib contrib-type="author" corresp="yes"><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Клименко</surname><given-names>Д. М.</given-names></name><name name-style="western" xml:lang="en"><surname>Klimenka</surname><given-names>D.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Клименко Д. М., студ.</p><p>220064, Минск, ул. Курчатова, 5;</p><p>Тел.: +375 17 209-58-36</p></bio><bio xml:lang="en"><p>Klimenka D., Studen</p><p>220064,  Minsk, Kurchatova St., 5;</p><p>Tel.: +375 17 209-58-36</p></bio><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Козлова</surname><given-names>Е. И.</given-names></name><name name-style="western" xml:lang="en"><surname>Kazlova</surname><given-names>А.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Козлова Е. И., канд. физ.-мат. наук, доц., зав. каф. интеллектуальных систем</p><p>220064, Минск, ул. Курчатова, 5;</p><p>Тел.: +375 17 209-58-36</p></bio><bio xml:lang="en"><p>Kazlova A., Cand. Sci. (Phys. and Math.) Associate Professor, Head of the Department of Intelligent Systems</p><p>220064,  Minsk, Kurchatova St., 5;</p><p>Tel.: +375 17 209-58-36</p><p> </p></bio><email xlink:type="simple">kozlova@bsu.by</email><xref ref-type="aff" rid="aff-1"/></contrib></contrib-group><aff-alternatives id="aff-1"><aff xml:lang="ru"><institution>Белорусский государственный университет</institution></aff><aff xml:lang="en"><institution>Belarusian State University</institution></aff></aff-alternatives><pub-date pub-type="collection"><year>2026</year></pub-date><pub-date pub-type="epub"><day>26</day><month>06</month><year>2026</year></pub-date><volume>32</volume><issue>2</issue><fpage>44</fpage><lpage>52</lpage><permissions><copyright-statement>Copyright &amp;#x00A9; Клименко Д.М., Козлова Е.И., 2026</copyright-statement><copyright-year>2026</copyright-year><copyright-holder xml:lang="ru">Клименко Д.М., Козлова Е.И.</copyright-holder><copyright-holder xml:lang="en">Klimenka D., Kazlova А.</copyright-holder><license xml:lang="ru" license-type="creative-commons-attribution" xlink:href="https://creativecommons.org/licenses/by/4.0/" xlink:type="simple"><license-p>Данная работа распространяется под лицензией Creative Commons Attribution 4.0.</license-p></license><license xml:lang="en" license-type="creative-commons-attribution" xlink:href="https://creativecommons.org/licenses/by/4.0/" xlink:type="simple"><license-p>This work is licensed under a Creative Commons Attribution 4.0 License.</license-p></license></permissions><self-uri xlink:href="https://dt.bsuir.by/jour/article/view/1039">https://dt.bsuir.by/jour/article/view/1039</self-uri><abstract><p>В настоящее время в области машинного обучения и искусственного интеллекта активно исследуются методы оптимизации и настройки гиперпараметров моделей. Одно из ключевых направлений таких исследований – анализ влияния изменения гиперпараметров, таких как количество слоев двумерной свертки (Conv2D) и их параметры (число фильтров, размер ядра), размер и шаг слоев максимальной подвыборки (MaxPooling2D), количество нейронов в полносвязных слоях, функции активации, размер пакета (batch_size) и количество эпох обучения, на точность предсказания модели. В статье приведен анализ влияния изменения количества слоев двумерной свертки, параметров слоев максимальной подвыборки (размер окна и шаг), количества нейронов в полносвязных слоях, выбора функции активации, размера пакета и числа эпох обучения на точность предсказания моделей машинного обучения на наборе данных Fashion-MNIST в архитектуре сверточной нейронной сети.</p></abstract><trans-abstract xml:lang="en"><p>Machine learning and artificial intelligence (AI) are currently actively researching methods for optimizing and tuning model hyperparameters. One key area of research is analyzing the impact of varying hyperparameters, such as the number of two-dimensional convolution (Conv2D) layers and their parameters (number of filters, kernel size), the size and stride of maximum pooling (MaxPooling2D) layers, the number of neurons in fully connected layers, activation functions, batch size (batch_size), and the number of training epochs, on the prediction accuracy of machine learning models using a convolutional neural network architecture on the Fashion-MNIST</p></trans-abstract><kwd-group xml:lang="ru"><kwd>размер пакета</kwd><kwd>двумерная свертка</kwd><kwd>Fashion-MNIST</kwd><kwd>максимальная подвыборка</kwd><kwd>гиперпараметры</kwd><kwd>количество эпох обучения</kwd><kwd>количество фильтров</kwd><kwd>оптимизация гиперпараметров</kwd><kwd>точность предсказания</kwd><kwd>функции активации</kwd></kwd-group><kwd-group xml:lang="en"><kwd>batch size</kwd><kwd>2D convolution</kwd><kwd>Fashion-MNIST</kwd><kwd>max pooling</kwd><kwd>hyperparameters</kwd><kwd>number of training epochs</kwd><kwd>number of filters</kwd><kwd>hyperparameter optimization</kwd><kwd>prediction accuracy</kwd><kwd>activation functions</kwd></kwd-group></article-meta></front><back><ref-list><title>References</title><ref id="cit1"><label>1</label><citation-alternatives><mixed-citation xml:lang="ru">Клейнер, С. Г. Исследование точности решения задачи оптимизации гиперпараметров с помощью нейронной сети [Электронный ресурс] / С. Г. Клейнер. Режим доступа: https://www.xn8sbempclcwd3bmt.xnp1ai/article/24427. Дата доступа: 24.12.2025.</mixed-citation><mixed-citation xml:lang="en">Kleiner, S. G. (2025) Study of the Accuracy of Solving a Hyper Parameter Optimization Problem Using a Neural Network. Available: https://www.xn8sbempclcwd3bmt.xnp1ai/article/24427 (Accessed 24 December 2025) (in Russian).</mixed-citation></citation-alternatives></ref><ref id="cit2"><label>2</label><citation-alternatives><mixed-citation xml:lang="ru">Игнатьева, С. А. Влияние выбора гиперпараметров при обучении cверточных нейронных сетей на точность повторной идентификации людей в системах видеонаблюдения [Электронный ресурс] / С. А. Игнатьева. Режим доступа: https://elib.psu.by/bitstream/123456789/34506/1/163-167.pdf. Дата доступа: 24.12.2025.</mixed-citation><mixed-citation xml:lang="en">Ignatieva S. A. (2025) The Impact of Hyper Parameter Selection in Convolutional Neural Network Training on the Accuracy of Person Re-Identification in Video Surveillance Systems. Available: https://elib.psu.by/bitstream/123456789/34506/1/163-167.pdf (Accessed 24 December 2025) (in Russian).</mixed-citation></citation-alternatives></ref><ref id="cit3"><label>3</label><citation-alternatives><mixed-citation xml:lang="ru">Fashion MNIST [Electronic resource]. Mode of access: https://github.com/zalandoresearch/fashion-mnist. Date of access: 24.12.2025.</mixed-citation><mixed-citation xml:lang="en">Fashion MNIST. Available: https://github.com/zalandoresearch/fashion-mnist (Accessed 24 December 2025) (in Russian).</mixed-citation></citation-alternatives></ref><ref id="cit4"><label>4</label><citation-alternatives><mixed-citation xml:lang="ru">Review of Deep Learning: Concepts, CNN Architectures, Challenges, Applications, Future Directions / L. Alzubaidi [et al.] // Journal of Big Data. 2021 Vol. 8, No 53 P. 2–74. https://doi.org/10.1186/s40537-021-00444-8.</mixed-citation><mixed-citation xml:lang="en">Alzubaidi L., Zhang J., Humaidi A. J., Al-Dujaili A., Duan Ye, Al-Shamma O., et al. (2021) Review of Deep Learning: Concepts, CNN Architectures, Challenges, Applications, Future Directions. Journal of Big Data. 8 (53), 2–74. https://doi.org/10.1186/s40537-021-00444-8.</mixed-citation></citation-alternatives></ref><ref id="cit5"><label>5</label><citation-alternatives><mixed-citation xml:lang="ru">Murphy, K. P. Probabilistic Machine Learning: An Introduction / К. Р. Murphy. USA: MIT Press, 2022 P. 323–358.</mixed-citation><mixed-citation xml:lang="en">Murphy K. P. (2022) Probabilistic Machine Learning: An Introduction. USA, MIT Press Publ. 323–358.</mixed-citation></citation-alternatives></ref></ref-list><fn-group><fn fn-type="conflict"><p>The authors declare that there are no conflicts of interest present.</p></fn></fn-group></back></article>
