Application of artificial neural networks in the problems of analysis of the dynamic structure of near-earth orbital space areas
The first experience of using artificial neural networks (ANNs) to study the dynamic structure of a selected region of near-Earth orbital space is described. The analysis of time series associated with the evolution of resonance characteristics that determine the dynamic structure of an area is usually performed manually. However in studying the dynamic structure of a large region of orbital space, the number of such time series is tens of thousands. As an alternative approach, the use of deep learning technologies is considered, namely the design of the architecture of a one-dimensional convolutional network and its training using the supervised learning method.
Download file
Counter downloads: 120
Keywords
численное моделирование, околоземные объекты, резонансы, искусственные нейронные сети, машинное обучение, numerical modeling, near-Earth objects, resonances, artificial neural networks, machine learningAuthors
| Name | Organization | |
| Krasavin D.S. | National Research Tomsk State University | iosfixed@gmail.com |
| Aleksandrova A.G. | National Research Tomsk State University | aleksandrovaannag@mail.ru |
| Tomilova I.V. | National Research Tomsk State University | irisha_tom@mail.ru |
References
Александрова А.Г., Бордовицына Т.В., Томилова И.В. // Астрон. вестн. - 2018. - Т. 52. - № 5. - С. 447-462.
Александрова А.Г., Бордовицына Т.В., Томилова И.В. // Изв. вузов. Физика. - 2018. - Т. 61. - № 4. - С. 75-80.
Александрова А.Г., Бордовицына Т.В., Александров В.Б. // Изв. вузов. Физика. - 2019. - Т. 62. - № 3. - С. 86-91.
Bishop C.M. Pattern Recognition and Machine Learning. - Springer, eBook, 2006. - 761 p.
Goodfellow I., Bengio Y., and Courville A. Deep Learning. - The MIT Press, eBook, 2016. - 800 p. - URL: http://www.deeplearningbook.org/contents/TOC.html (06.12.2019).
Описание библиотеки torch для python. - URL: https://github.com/pytorch/pytorch (06.12.2019).
Описание пакет nn библиотеки torch для языка python. - URL: https://pytorch.org/docs/stable/nn.html (06.12.2019).
Плас Дж. Вандер. Python для сложных задач: наука о данных и машинное обучение. - СПб.: Питер, 2018. - 576 с.
Рашка С. Python и машинное обучение: пер. с англ. - М.: ДМК Пресс, 2017. - 420 c.
Александрова А.Г., Бордовицына Т.В., Чувашов И.Н. // Изв. вузов. Физика. - 2017. - Т. 60. - № 1. - С. 69-76.
Ismail Fawaz H., Forestier G., Weber J., et al. // Data Mining and Knowledge Discovery. - 2019. - V. 33. - Iss. 4. - P. 917-963. DOI: 10.1007/s10618-019-00619-1.
Hagan M.T., Demuth H.B., Hudson Beale M., and Jesús O. Neural Network Design. - 2nd Edition. - eBook, 2019. - 1012 p. - URL: https://hagan.okstate.edu/nnd.html (06.12.2019).
Томилова И.В., Красавин Д.С., Бордовицына Т.В. // Астрон. вестн. - 2020. - Т. 54 (в печати).
