<?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">ntv</journal-id><journal-title-group><journal-title xml:lang="ru">Научно-технический вестник информационных технологий, механики и оптики</journal-title><trans-title-group xml:lang="en"><trans-title>Scientific and Technical Journal of Information Technologies, Mechanics and Optics</trans-title></trans-title-group></journal-title-group><issn pub-type="ppub">2226-1494</issn><issn pub-type="epub">2500-0373</issn><publisher><publisher-name>Университет ИТМО</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="doi">10.17586/2226-1494-2024-24-4-608-614</article-id><article-id custom-type="elpub" pub-id-type="custom">ntv-310</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 И КОМПЬЮТЕРНОЕ МОДЕЛИРОВАНИЕ</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="en"><subject>MODELING AND SIMULATION</subject></subj-group></article-categories><title-group><article-title>Моделирование многомерных данных с помощью композитных байесовских сетей</article-title><trans-title-group xml:lang="en"><trans-title>Flexible and tractable modeling of multivariate data using composite Bayesian networks</trans-title></trans-title-group></title-group><contrib-group><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0001-8679-5868</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Деева</surname><given-names>И. Ю.</given-names></name><name name-style="western" xml:lang="en"><surname>Deeva</surname><given-names>I. Yu.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Деева Ирина Юрьевна — кандидат физико-математических наук, старший научный сотрудник</p><p>Санкт-Петербург, 197101</p></bio><bio xml:lang="en"><p>Irina Yu. Deeva — PhD (Physics &amp; Mathematics), Senior Researcher</p><p>Saint Petersburg, 197101</p></bio><email xlink:type="simple">iriny.deeva@gmail.com</email><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0009-0003-2606-431X</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Шахкян</surname><given-names>К. А.</given-names></name><name name-style="western" xml:lang="en"><surname>Shakhkyan</surname><given-names>K. A.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Шахкян Каринэ Артуровна — инженер</p><p>Санкт-Петербург, 197101</p></bio><bio xml:lang="en"><p>Karine A. Shakhkyan — Engineer</p><p>Saint Petersburg, 197101</p></bio><email xlink:type="simple">kshahkyan@yandex.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/0009-0006-6418-6117</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Каминский</surname><given-names>Ю. К.</given-names></name><name name-style="western" xml:lang="en"><surname>Kaminsky</surname><given-names>Yu. K.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Каминский Юрий Константинович — инженер</p><p>Санкт-Петербург, 197101</p></bio><bio xml:lang="en"><p>Yury K. Kaminsky — Engineer</p><p>Saint Petersburg, 197101</p></bio><email xlink:type="simple">jkaminski@niuitmo.ru</email><xref ref-type="aff" rid="aff-1"/></contrib></contrib-group><aff-alternatives id="aff-1"><aff xml:lang="ru"><institution>Университет ИТМО</institution><country>Россия</country></aff><aff xml:lang="en"><institution>ITMO University</institution><country>Russian Federation</country></aff></aff-alternatives><pub-date pub-type="collection"><year>2024</year></pub-date><pub-date pub-type="epub"><day>11</day><month>12</month><year>2024</year></pub-date><volume>24</volume><issue>4</issue><fpage>608</fpage><lpage>614</lpage><permissions><copyright-statement>Copyright &amp;#x00A9; Деева И.Ю., Шахкян К.А., Каминский Ю.К., 2024</copyright-statement><copyright-year>2024</copyright-year><copyright-holder xml:lang="ru">Деева И.Ю., Шахкян К.А., Каминский Ю.К.</copyright-holder><copyright-holder xml:lang="en">Deeva I.Y., Shakhkyan K.A., Kaminsky Y.K.</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://ntv.elpub.ru/jour/article/view/310">https://ntv.elpub.ru/jour/article/view/310</self-uri><abstract><p>Введение. Представлен новый подход к моделированию нелинейных зависимостей, названный композитными байесовскими сетями. Основной акцент сделан на интеграции моделей машинного обучения в байесовские сети с сохранением их основополагающих принципов. Новизна предложенного подхода состоит в том, что он позволяет решить проблему несоответствия данных традиционным предположениям о зависимостях. Метод. Представленный подход заключается в подборе разнообразных моделей машинного обучения на этапе обучения композитных байесовских сетей. Это позволяет гибко настраивать характер зависимостей в соответствии с требованиями и продиктованными характеристиками моделируемого объекта. Программная реализация подхода выполнена в виде специализированного фреймворка, описывающего все необходимые функциональные возможности. Основные результаты. Проведена экспериментальная оценка эффективности моделирования зависимостей между признаками. Для экспериментов выбраны для бенчмарков и из репозитория UCI для реальных данных. Эффективность предложенных композитных байесовских сетей подтверждена сравнением значений правдоподобия и показателя F1 с классическими байесовскими сетями, обученными алгоритмом Hill-Climbing. Показана высокая точность представления многомерных распределений. При этом на бенчмарках улучшение оказалось незначительным, поскольку они содержат линейные зависимости, которые хорошо моделируются классическими алгоритмами. На реальных наборах данных UCI получено улучшение правдоподобия в среднем на 30 %. Обсуждение. Полученные результаты могут найти применение в областях, требующих моделирования сложных зависимостей между признаками, например, в машинном обучении, статистике, задачах анализа данных, а также в конкретных предметных областях.</p></abstract><trans-abstract xml:lang="en"><p>The article presents a new approach to modeling nonlinear dependencies called composite Bayesian networks. The main emphasis is on integrating machine learning models into Bayesian networks while maintaining their fundamental principles. The novelty of the approach is that it allows us to solve the problem of data inconsistency with traditional assumptions about dependencies. The presented method consists in selecting a variety of machine learning models at the stage of training composite Bayesian networks. This allows you to flexibly customize the nature of the dependencies in accordance with the requirements and dictated characteristics of the modeled object. The software implementation is made in the form of a specialized framework that describes all the necessary functionality. The results of experiments to evaluate the effectiveness of modeling dependencies between features are presented. Data for the experiments was taken from the bnlearn repository for benchmarks and from the UCI repository for real data. The performance of composite Bayesian networks was validated by comparing the likelihood and F1 score with classical Bayesian networks trained with the Hill-Climbing algorithm, demonstrating high accuracy in representing multivariate distributions. The improvement in benchmarks is insignificant since they contain linear dependencies that are well modeled by the classical algorithm. An average 30 % improvement in likelihood was obtained on real UCI datasets. The obtained data can be applied in areas that require modeling complex dependencies between features, for example, in machine learning, statistics, data analysis, as well as in specific subject areas.</p></trans-abstract><kwd-group xml:lang="ru"><kwd>байесовские сети</kwd><kwd>вероятностные графовые модели</kwd><kwd>обучение параметров</kwd><kwd>модели машинного обучения</kwd><kwd>генетический алгоритм</kwd></kwd-group><kwd-group xml:lang="en"><kwd>Bayesian networks</kwd><kwd>probabilistic graph models</kwd><kwd>parameter learning</kwd><kwd>machine learning models</kwd><kwd>genetic algorithm</kwd></kwd-group><funding-group><funding-statement xml:lang="ru">Работа выполнена в рамках государственного задания Министерства науки и высшего образования Российской Федерации (проект № FSER-2024-0004).</funding-statement><funding-statement xml:lang="en">The research was carried out within the state assignment of the Ministry of Science and Higher Education of the Russian Federation (project No. FSER-2024-0004).</funding-statement></funding-group></article-meta></front><back><ref-list><title>References</title><ref id="cit1"><label>1</label><citation-alternatives><mixed-citation xml:lang="ru">Handbook of Graphical Models / ed. by M. Maathuis, M. Drton, S. Lauritzen, M. Wainwright. CRC Press, 2018. 554 p. https://doi.org/10.1201/9780429463976</mixed-citation><mixed-citation xml:lang="en">Handbook of Graphical Models. Ed. by M. Maathuis, M. Drton, S. Lauritzen, M. Wainwright. CRC Press, 2018, 554 p. https://doi.org/10.1201/9780429463976</mixed-citation></citation-alternatives></ref><ref id="cit2"><label>2</label><citation-alternatives><mixed-citation xml:lang="ru">Mascaro S., Nicholso A.E., Korb K.B. Anomaly detection in vessel tracks using Bayesian networks // International Journal of Approximate Reasoning. 2014. V. 55. N 1. P. 84–98. https://doi.org/10.1016/j.ijar.2013.03.012</mixed-citation><mixed-citation xml:lang="en">Mascaro S., Nicholso A.E., Korb K.B. Anomaly detection in vessel tracks using Bayesian networks. International Journal of Approximate Reasoning, 2014, vol. 55, no. 1, pp. 84–98. https://doi.org/10.1016/j.ijar.2013.03.012</mixed-citation></citation-alternatives></ref><ref id="cit3"><label>3</label><citation-alternatives><mixed-citation xml:lang="ru">McLachlan S., Dube K., Hitman G.A., Fenton N.E., Kyrimi E. Bayesian networks in healthcare: Distribution by medical condition // Artificial Intelligence in Medicine. 2020. V. 107. P. 101912. https://doi.org/10.1016/j.artmed.2020.101912</mixed-citation><mixed-citation xml:lang="en">McLachlan S., Dube K., Hitman G.A., Fenton N.E., Kyrimi E. Bayesian networks in healthcare: Distribution by medical condition. Artificial Intelligence in Medicine, 2020, vol. 107, pp. 101912. https://doi.org/10.1016/j.artmed.2020.101912</mixed-citation></citation-alternatives></ref><ref id="cit4"><label>4</label><citation-alternatives><mixed-citation xml:lang="ru">Friedman N., Goldszmidt M. Learning Bayesian networks with local structure // NATO ASI Series. 1998. V. 89. P. 421–459. https://doi.org/10.1007/978-94-011-5014-9_15</mixed-citation><mixed-citation xml:lang="en">Friedman N., Goldszmidt M. Learning Bayesian networks with local structure. NATO ASI Series, 1998, vol. 89, pp. 421–459. https://doi.org/10.1007/978-94-011-5014-9_15</mixed-citation></citation-alternatives></ref><ref id="cit5"><label>5</label><citation-alternatives><mixed-citation xml:lang="ru">Grzegorczyk M. An introduction to gaussian bayesian networks // Methods in Molecular Biology. 2010. V. 662. P. 121–147. https://doi.org/10.1007/978-1-60761-800-3_6</mixed-citation><mixed-citation xml:lang="en">Grzegorczyk M. An introduction to gaussian bayesian networks. Methods in Molecular Biology, 2010, vol. 662, pp. 121–147. https://doi.org/10.1007/978-1-60761-800-3_6</mixed-citation></citation-alternatives></ref><ref id="cit6"><label>6</label><citation-alternatives><mixed-citation xml:lang="ru">Lerner U., Segal E., Koller D. Exact inference in networks with discrete children of continuous parents // arXiv. 2013. arXiv:1301.2289. https://doi.org/10.48550/arXiv.1301.2289</mixed-citation><mixed-citation xml:lang="en">Lerner U., Segal E., Koller D. Exact inference in networks with discrete children of continuous parents. arXiv, 2013, arXiv:1301.2289. https://doi.org/10.48550/arXiv.1301.2289</mixed-citation></citation-alternatives></ref><ref id="cit7"><label>7</label><citation-alternatives><mixed-citation xml:lang="ru">Pérez A., Larrañaga P., Inza I. Bayesian classifiers based on kernel density estimation: Flexible classifiers // International Journal of Approximate Reasoning. 2009. V. 50. N 2. P. 341–362. https://doi.org/10.1016/j.ijar.2008.08.008</mixed-citation><mixed-citation xml:lang="en">Pérez A., Larrañaga P., Inza I. Bayesian classifiers based on kernel density estimation: Flexible classifiers. International Journal of Approximate Reasoning, 2009, vol. 50, no. 2, pp. 341–362. https://doi.org/10.1016/j.ijar.2008.08.008</mixed-citation></citation-alternatives></ref><ref id="cit8"><label>8</label><citation-alternatives><mixed-citation xml:lang="ru">Ickstadt K., Bornkamp B., Grzegorczyk M., Wieczorek J., Sheriff M.R., Grecco H.E., Zamir E. Nonparametric Bayesian networks // Bayesian Statistics 9. 2011. P. 283–316. https://doi.org/10.1093/acprof:oso/9780199694587.003.0010</mixed-citation><mixed-citation xml:lang="en">Ickstadt K., Bornkamp B., Grzegorczyk M., Wieczorek J., Sheriff M.R., Grecco H.E., Zamir E. Nonparametric Bayesian networks. Bayesian Statistics 9, 2011, pp. 283–316. https://doi.org/10.1093/acprof:oso/9780199694587.003.0010</mixed-citation></citation-alternatives></ref><ref id="cit9"><label>9</label><citation-alternatives><mixed-citation xml:lang="ru">Deeva I., Bubnova A., Kalyuzhnaya A.V. Advanced approach for distributions parameters learning in Bayesian networks with gaussian mixture models and discriminative models // Mathematics. 2023. V. 11. N 2. P. 343. https://doi.org/10.3390/math11020343</mixed-citation><mixed-citation xml:lang="en">Deeva I., Bubnova A., Kalyuzhnaya A.V. Advanced approach for distributions parameters learning in Bayesian networks with gaussian mixture models and discriminative models. Mathematics, 2023, vol. 11, no. 2, pp. 343. https://doi.org/10.3390/math11020343</mixed-citation></citation-alternatives></ref><ref id="cit10"><label>10</label><citation-alternatives><mixed-citation xml:lang="ru">Langseth H., Nielsen T.D., Rumí R., Salmerón A. Mixtures of truncated basis functions // International Journal of Approximate Reasoning. 2012. V. 53. N 2. P. 212–227. https://doi.org/10.1016/j.ijar.2011.10.004</mixed-citation><mixed-citation xml:lang="en">Langseth H., Nielsen T.D., Rumí R., Salmerón A. Mixtures of truncated basis functions. International Journal of Approximate Reasoning, 2012, vol. 53, no. 2, pp. 212–227. https://doi.org/10.1016/j.ijar.2011.10.004</mixed-citation></citation-alternatives></ref><ref id="cit11"><label>11</label><citation-alternatives><mixed-citation xml:lang="ru">Atienza D., Larrañaga P., Bielza C. Hybrid semiparametric Bayesian networks // TEST. 2022. V. 31. N 2. P. 299–327. https://doi.org/10.1007/s11749-022-00812-3</mixed-citation><mixed-citation xml:lang="en">Atienza D., Larrañaga P., Bielza C. Hybrid semiparametric Bayesian networks. TEST, 2022, vol. 31, no. 2, pp. 299–327. https://doi.org/10.1007/s11749-022-00812-3</mixed-citation></citation-alternatives></ref><ref id="cit12"><label>12</label><citation-alternatives><mixed-citation xml:lang="ru">Sloman S. Causal Models: How People Think about the World and Its Alternatives. Oxford University Press, 2005. 211 p. https://doi.org/10.1093/acprof:oso/9780195183115.001.0001</mixed-citation><mixed-citation xml:lang="en">Sloman S. Causal Models: How People Think about the World and Its Alternatives. Oxford University Press, 2005, 211 p. https://doi.org/10.1093/acprof:oso/9780195183115.001.0001</mixed-citation></citation-alternatives></ref><ref id="cit13"><label>13</label><citation-alternatives><mixed-citation xml:lang="ru">Larrañaga P., Karshenas H., Bielza C., Santana R. A review on evolutionary algorithms in Bayesian network learning and inference tasks // Information Sciences. 2013. V. 233. P. 109–125. https://doi.org/10.1016/j.ins.2012.12.051</mixed-citation><mixed-citation xml:lang="en">Larrañaga P., Karshenas H., Bielza C., Santana R. A review on evolutionary algorithms in Bayesian network learning and inference tasks. Information Sciences, 2013, vol. 233, pp. 109–125. https://doi.org/10.1016/j.ins.2012.12.051</mixed-citation></citation-alternatives></ref><ref id="cit14"><label>14</label><citation-alternatives><mixed-citation xml:lang="ru">Gámez J.A., Mateo J.L., Puerta J.M. Learning Bayesian networks by hill climbing: efficient methods based on progressive restriction of the neighborhood // Data Mining and Knowledge Discovery. 2011. V. 22. N 1-2. P. 106–148. https://doi.org/10.1007/s10618-010-0178-6</mixed-citation><mixed-citation xml:lang="en">Gámez J.A., Mateo J.L., Puerta J.M. Learning Bayesian networks by hill climbing: efficient methods based on progressive restriction of the neighborhood. Data Mining and Knowledge Discovery, 2011, vol. 22, no. 1-2, pp. 106–148. https://doi.org/10.1007/s10618-010-0178-6</mixed-citation></citation-alternatives></ref><ref id="cit15"><label>15</label><citation-alternatives><mixed-citation xml:lang="ru">Behjati S., Beigy H. Improved K2 algorithm for Bayesian network structure learning // Engineering Applications of Artificial Intelligence. 2020. V. 91. P. 103617. https://doi.org/10.1016/j.engappai.2020.103617</mixed-citation><mixed-citation xml:lang="en">Behjati S., Beigy H. Improved K2 algorithm for Bayesian network structure learning. Engineering Applications of Artificial Intelligence, 2020, vol. 91, pp. 103617. https://doi.org/10.1016/j.engappai.2020.103617</mixed-citation></citation-alternatives></ref><ref id="cit16"><label>16</label><citation-alternatives><mixed-citation xml:lang="ru">Lerner B., Malka R. Investigation of the K2 algorithm in learning Bayesian network classifiers // Applied Artificial Intelligence. 2011. V. 25. N 1. P. 74–96. https://doi.org/10.1080/08839514.2011.529265</mixed-citation><mixed-citation xml:lang="en">Lerner B., Malka R. Investigation of the K2 algorithm in learning Bayesian network classifiers. Applied Artificial Intelligence, 2011, vol. 25, no. 1, pp. 74–96. https://doi.org/10.1080/08839514.2011.529265</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>
