<?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">mireabulletin</journal-id><journal-title-group><journal-title xml:lang="ru">Russian Technological Journal</journal-title><trans-title-group xml:lang="en"><trans-title>Russian Technological Journal</trans-title></trans-title-group></journal-title-group><issn pub-type="ppub">2782-3210</issn><issn pub-type="epub">2500-316X</issn><publisher><publisher-name>RTU MIREA</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="doi">10.32362/2500-316X-2026-14-1-82-90</article-id><article-id custom-type="edn" pub-id-type="custom">KCUBAG</article-id><article-id custom-type="elpub" pub-id-type="custom">mireabulletin-1367</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>МАТЕМАТИЧЕСКОЕ МОДЕЛИРОВАНИЕ</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="en"><subject>MATHEMATICAL MODELING</subject></subj-group></article-categories><title-group><article-title>Трансформация пространства признаков в методе опорных векторов</article-title><trans-title-group xml:lang="en"><trans-title>Feature space transformation in the support vector method</trans-title></trans-title-group></title-group><contrib-group><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0009-0003-2314-7400</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Федоров</surname><given-names>А. В.</given-names></name><name name-style="western" xml:lang="en"><surname>Fedorov</surname><given-names>A. V.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Федоров Алексей Викторович - аспирант, кафедра высшей математики, Институт искусственного интеллекта.</p><p>119454, Москва, пр-т Вернадского, д. 78</p></bio><bio xml:lang="en"><p>Aleksey V. Fedorov - Postgraduate Student, Higher Mathematics Department, Institute of Artificial Intelligence, MIREA – Russian Technological University.</p><p>78, Vernadskogo pr., Moscow, 119454</p></bio><email xlink:type="simple">fedorov_av@mirea.ru</email><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0009-0004-0905-3827</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Парфенов</surname><given-names>Д. В.</given-names></name><name name-style="western" xml:lang="en"><surname>Parfenov</surname><given-names>D. V.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Парфенов Денис Васильевич - к.т.н., доцент, кафедра высшей математики, Институт искусственного интеллекта.</p><p>119454, Москва, пр-т Вернадского, д. 78</p><p>Scopus Author ID 57217119805</p></bio><bio xml:lang="en"><p>Denis V. Parfenov - Cand. Sci. (Eng.), Associate Professor, Higher Mathematics Department, Institute of Artificial Intelligence, MIREA – Russian Technological University.</p><p>78, Vernadskogo pr., Moscow, 119454</p><p>Scopus Author ID 57217119805</p></bio><email xlink:type="simple">parfenov@mirea.ru</email><xref ref-type="aff" rid="aff-1"/></contrib></contrib-group><aff-alternatives id="aff-1"><aff xml:lang="ru"><institution>МИРЭА – Российский технологический университет</institution><country>Россия</country></aff><aff xml:lang="en"><institution>MIREA – Russian Technological University</institution><country>Russian Federation</country></aff></aff-alternatives><pub-date pub-type="collection"><year>2026</year></pub-date><pub-date pub-type="epub"><day>05</day><month>02</month><year>2026</year></pub-date><volume>14</volume><issue>1</issue><fpage>82</fpage><lpage>90</lpage><permissions><copyright-statement>Copyright &amp;#x00A9; Федоров А.В., Парфенов Д.В., 2026</copyright-statement><copyright-year>2026</copyright-year><copyright-holder xml:lang="ru">Федоров А.В., Парфенов Д.В.</copyright-holder><copyright-holder xml:lang="en">Fedorov A.V., Parfenov D.V.</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://www.rtj-mirea.ru/jour/article/view/1367">https://www.rtj-mirea.ru/jour/article/view/1367</self-uri><abstract><sec><title>Цели</title><p>Цели. Работа посвящена разработке и исследованию обобщенного нелинейного метода опорных векторов (support vector machine, SVM) с использованием адаптивной трансформации пространства признаков, направленного на улучшение вычислительной эффективности при сохранении высокого качества классификации. В качестве задачи-примера рассматривается двухклассовая классификация. Целью исследования является количественная оценка производительности предложенного подхода в сравнении с классическими SVM-моделями, использующими фиксированные ядровые функции, а также изучение влияния параметров трансформации на качество классификации.</p></sec><sec><title>Методы</title><p>Методы. Предлагается модифицированный подход, при котором входные данные предварительно преобразуются с помощью обучаемого нелинейного отображения фиксированной структуры. Это отображение реализуется в виде композиции элементарных функций и параметризуется ограниченным числом обучаемых весов, что обеспечивает контроль над сложностью модели. После трансформации применяется линейный SVM с L2-регуляризацией. Для обучения модели используются стандартные методы численной оптимизации без ограничений. Качество классификации оценивается с помощью метрики точности (Accuracy), усредненной по результатам 10-кратной перекрестной валидации. Рассматривается поведение модели при изменении размерности признакового пространства. Проводится анализ вычислительной сложности по числу операций и времени применения модели на тестовых выборках.</p></sec><sec><title>Результаты</title><p>Результаты. Численные эксперименты показали, что предложенная модель позволяет существенно сократить время классификации по сравнению с SVM с полиномиальным ядром, обеспечивая при этом сопоставимое качество. Анализ временных затрат подтвердил, что предложенный подход масштабируется значительно лучше, чем классические ядровые методы. При этом структура модели сохраняет интерпретируемость и может быть дополнительно адаптирована под особенности предметной области.</p></sec><sec><title>Выводы</title><p>Выводы. Разработанный метод представляет собой эффективную альтернативу классическим ядровым алгоритмам. Благодаря параметризуемому отображению признакового пространства он обеспечивает адаптивность, интерпретируемость и масштабируемость, что делает его перспективным для практического применения в задачах машинного обучения.</p></sec></abstract><trans-abstract xml:lang="en"><sec><title>Objectives</title><p>Objectives. This study focuses on the development and investigation of a generalized nonlinear Support Vector Machine (SVM) method incorporating an adaptive transformation of the feature space. Its aim is to improve computational efficiency while maintaining high classification accuracy. The binary classification problem is used as a case study. The main objective of the research is to quantitatively evaluate the performance of the proposed approach when compared to classical SVM models using fixed kernel functions, and to analyze how the transformation parameters affect classification quality.</p></sec><sec><title>Methods</title><p>Methods. The proposed approach involves a preliminary transformation of the input data using a learnable nonlinear mapping with a fixed structure. This mapping is implemented as a composition of elementary functions and is parameterized by a limited number of trainable weights which allows control over model complexity. A linear SVM with L2 regularization is applied after the transformation. The model is trained using conventional, unconstrained numerical optimization methods. The classification quality is evaluated using the Accuracy metric averaged over 10-fold cross-validation. The work also studies the behavior of the model with varying feature space dimensionality. In addition, computational complexity is analyzed in terms of the number of operations and inference time required on test datasets.</p></sec><sec><title>Results</title><p>Results. Numerical experiments demonstrate that the proposed model significantly reduces classification time when compared to a polynomial-kernel SVM, while maintaining a comparable level of accuracy. The runtime analysis confirms that the proposed approach scales much better than traditional kernel methods. At the same time, the structure of the model remains interpretable and can be further adapted to the specifics of the application domain.</p></sec><sec><title>Conclusions</title><p>Conclusions. The method developed provides an efficient alternative to traditional kernel-based algorithms. Through the use of a parameterized transformation of the feature space, the method enables adaptability, interpretability, and scalability, making it promising for practical applications in machine learning tasks.</p></sec></trans-abstract><kwd-group xml:lang="ru"><kwd>классификация</kwd><kwd>преобразование признакового пространства</kwd><kwd>нелинейное отображение</kwd><kwd>нелинейный метод опорных векторов</kwd><kwd>функции ядра</kwd><kwd>вычислительная сложность</kwd></kwd-group><kwd-group xml:lang="en"><kwd>classification</kwd><kwd>feature space transformation</kwd><kwd>nonlinear mapping</kwd><kwd>nonlinear support vector machine</kwd><kwd>kernel functions</kwd><kwd>computational complexity</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">Vapnik V. Statistical Learning Theory. New York: Wiley; 1998, 736 p. ISBN 978-0-471-03003-4</mixed-citation><mixed-citation xml:lang="en">Vapnik V. Statistical Learning Theory. New York: Wiley; 1998, 736 p. ISBN 978-0-471-03003-4</mixed-citation></citation-alternatives></ref><ref id="cit2"><label>2</label><citation-alternatives><mixed-citation xml:lang="ru">Maggioni F., Spinelli A. A novel robust optimization model for nonlinear Support Vector Machine. Computers &amp; Operations Research. 2024;157:105059.</mixed-citation><mixed-citation xml:lang="en">Maggioni F., Spinelli A. A novel robust optimization model for nonlinear Support Vector Machine. Computers &amp; Operations Research. 2024;157:105059.</mixed-citation></citation-alternatives></ref><ref id="cit3"><label>3</label><citation-alternatives><mixed-citation xml:lang="ru">Rubin N., Fischer K., Lindner J., Dahmen D., Seroussi I., Ringel Z., Krüger M., Helias M. From Kernels to Features: A Multi-Scale Adaptive Theory of Feature Learning. arXiv preprint arXiv:2402.03210. 2024. https://doi.org/10.48550/arXiv.2502.03210</mixed-citation><mixed-citation xml:lang="en">Rubin N., Fischer K., Lindner J., Dahmen D., Seroussi I., Ringel Z., Krüger M., Helias M. From Kernels to Features: A Multi-Scale Adaptive Theory of Feature Learning. arXiv preprint arXiv:2402.03210. 2024. https://doi.org/10.48550/arXiv.2502.03210</mixed-citation></citation-alternatives></ref><ref id="cit4"><label>4</label><citation-alternatives><mixed-citation xml:lang="ru">LeJeune D., Alemohammad S. An Adaptive Tangent Feature Perspective of Neural Networks. In: Proceedings of the 37th International Conference on Machine Learning (ICML). 2024. URL: https://proceedings.mlr.press/v234/lejeune24a/lejeune24a.pdf. Дата обращения 07.07.2025. / Accessed July 07, 2025.</mixed-citation><mixed-citation xml:lang="en">LeJeune D., Alemohammad S. An Adaptive Tangent Feature Perspective of Neural Networks. In: Proceedings of the 37th International Conference on Machine Learning (ICML). 2024. URL: https://proceedings.mlr.press/v234/lejeune24a/lejeune24a.pdf. Дата обращения 07.07.2025. / Accessed July 07, 2025.</mixed-citation></citation-alternatives></ref><ref id="cit5"><label>5</label><citation-alternatives><mixed-citation xml:lang="ru">Li Y., Lin Q. Diagonal Over-parameterization in Reproducing Kernel Hilbert Spaces as an Adaptive Feature Model: Generalization and Adaptivity. arXiv preprint arXiv:2501.08679. 2025. https://arxiv.org/abs/2501.08679</mixed-citation><mixed-citation xml:lang="en">Li Y., Lin Q. Diagonal Over-parameterization in Reproducing Kernel Hilbert Spaces as an Adaptive Feature Model: Generalization and Adaptivity. arXiv preprint arXiv:2501.08679. 2025. https://arxiv.org/abs/2501.08679</mixed-citation></citation-alternatives></ref><ref id="cit6"><label>6</label><citation-alternatives><mixed-citation xml:lang="ru">Kurbucz M.T., Benkő Z., Varga L., et al. Adaptive Law-Based Transformation (ALT): A Fast and Transparent Feature Transformation Method for Time Series Classification. arXiv preprint arXiv:2501.09217. 2025. https://arxiv.org/abs/2501.09217</mixed-citation><mixed-citation xml:lang="en">Kurbucz M.T., Benkő Z., Varga L., et al. Adaptive Law-Based Transformation (ALT): A Fast and Transparent Feature Transformation Method for Time Series Classification. arXiv preprint arXiv:2501.09217. 2025. https://arxiv.org/abs/2501.09217</mixed-citation></citation-alternatives></ref><ref id="cit7"><label>7</label><citation-alternatives><mixed-citation xml:lang="ru">Tikhonov A.N., Arsenin V.Y. Solutions of Ill-Posed Problems. Washington, D.C.: W.H. Freeman and Co.; 1977, 258 p.</mixed-citation><mixed-citation xml:lang="en">Tikhonov A.N., Arsenin V.Y. Solutions of Ill-Posed Problems. Washington, D.C.: W.H. Freeman and Co.; 1977, 258 p.</mixed-citation></citation-alternatives></ref><ref id="cit8"><label>8</label><citation-alternatives><mixed-citation xml:lang="ru">Bishop C.M. Pattern Recognition and Machine Learning. New York: Springer; 2006, 738 p.</mixed-citation><mixed-citation xml:lang="en">Bishop C.M. Pattern Recognition and Machine Learning. New York: Springer; 2006, 738 p.</mixed-citation></citation-alternatives></ref><ref id="cit9"><label>9</label><citation-alternatives><mixed-citation xml:lang="ru">Fletcher R. Practical Methods of Optimization. New York: John Wiley &amp; Sons; 1987, 464 p.</mixed-citation><mixed-citation xml:lang="en">Fletcher R. Practical Methods of Optimization. New York: John Wiley &amp; Sons; 1987, 464 p.</mixed-citation></citation-alternatives></ref><ref id="cit10"><label>10</label><citation-alternatives><mixed-citation xml:lang="ru">Baydin A.G., Pearlmutter B.A., Radul A.A., Siskind J.M. Automatic differentiation in machine learning: a survey. Journal of Machine Learning Research (JMLR). 2018;18(153):1–43.</mixed-citation><mixed-citation xml:lang="en">Baydin A.G., Pearlmutter B.A., Radul A.A., Siskind J.M. Automatic differentiation in machine learning: a survey. Journal of Machine Learning Research (JMLR). 2018;18(153):1–43.</mixed-citation></citation-alternatives></ref><ref id="cit11"><label>11</label><citation-alternatives><mixed-citation xml:lang="ru">Fisher J. Automatic differentiation with dual numbers. arXiv preprint arXiv:2201.00024.</mixed-citation><mixed-citation xml:lang="en">Fisher J. Automatic differentiation with dual numbers. arXiv preprint arXiv:2201.00024.</mixed-citation></citation-alternatives></ref><ref id="cit12"><label>12</label><citation-alternatives><mixed-citation xml:lang="ru">Bischl B., Casalicchio G., Feurer M., Hutter F., Lang M., Mantovani R.G., van Rijn J.N., Vanschoren J. OpenML Benchmarking Suites. arXiv preprint arXiv:1708.03731v2 [stat.ML], 2019. https://arxiv.org/abs/1708.03731v2</mixed-citation><mixed-citation xml:lang="en">Bischl B., Casalicchio G., Feurer M., Hutter F., Lang M., Mantovani R.G., van Rijn J.N., Vanschoren J. OpenML Benchmarking Suites. arXiv preprint arXiv:1708.03731v2 [stat.ML], 2019. https://arxiv.org/abs/1708.03731v2</mixed-citation></citation-alternatives></ref><ref id="cit13"><label>13</label><citation-alternatives><mixed-citation xml:lang="ru">Kohavi R. A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection. In: Proceedings of Fourteenth International Joint Conference on Artificial Intelligence (IJCAI). 1995;14(2):1137–1143.</mixed-citation><mixed-citation xml:lang="en">Kohavi R. A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection. In: Proceedings of Fourteenth International Joint Conference on Artificial Intelligence (IJCAI). 1995;14(2):1137–1143.</mixed-citation></citation-alternatives></ref><ref id="cit14"><label>14</label><citation-alternatives><mixed-citation xml:lang="ru">Labatut V., Cherifi H. Accuracy Measures for the Comparison of Classifiers. arXiv preprint arXiv:1207.3790. https://doi.org/10.48550/arXiv.1207.3790</mixed-citation><mixed-citation xml:lang="en">Labatut V., Cherifi H. Accuracy Measures for the Comparison of Classifiers. arXiv preprint arXiv:1207.3790. https://doi.org/10.48550/arXiv.1207.3790</mixed-citation></citation-alternatives></ref><ref id="cit15"><label>15</label><citation-alternatives><mixed-citation xml:lang="ru">Schölkopf B., Smola A.J. Learning with Kernels. Cambridge: MIT Press; 2002, 300 p.</mixed-citation><mixed-citation xml:lang="en">Schölkopf B., Smola A.J. Learning with Kernels. Cambridge: MIT Press; 2002, 300 p.</mixed-citation></citation-alternatives></ref><ref id="cit16"><label>16</label><citation-alternatives><mixed-citation xml:lang="ru">Hofmann T., Schölkopf B., Smola A.J. Kernel Methods in Machine Learning. Ann. Statist. 2008;36(3):1171–1220. https://doi.org/10.1214/009053607000000677</mixed-citation><mixed-citation xml:lang="en">Hofmann T., Schölkopf B., Smola A.J. Kernel Methods in Machine Learning. Ann. Statist. 2008;36(3):1171–1220. https://doi.org/10.1214/009053607000000677</mixed-citation></citation-alternatives></ref><ref id="cit17"><label>17</label><citation-alternatives><mixed-citation xml:lang="ru">Boser B.E., Guyon I.M., Vapnik V. A Training Algorithm for Optimal Margin Classifiers. In: Proceedings of the Fifth Annual Workshop on Computational Learning Theory (COLT). 1992. P. 144–152. https://doi.org/10.1145/130385.130401</mixed-citation><mixed-citation xml:lang="en">Boser B.E., Guyon I.M., Vapnik V. A Training Algorithm for Optimal Margin Classifiers. In: Proceedings of the Fifth Annual Workshop on Computational Learning Theory (COLT). 1992. P. 144–152. https://doi.org/10.1145/130385.130401</mixed-citation></citation-alternatives></ref><ref id="cit18"><label>18</label><citation-alternatives><mixed-citation xml:lang="ru">Cortes C., Vapnik V. Support-Vector Networks. Mach. Learn. 1995;20(3):273–297. https://doi.org/10.1007/BF00994018</mixed-citation><mixed-citation xml:lang="en">Cortes C., Vapnik V. Support-Vector Networks. Mach. Learn. 1995;20(3):273–297. https://doi.org/10.1007/BF00994018</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>
