Фейгин А.М.
Мухин Д.Н.
Лоскутов Е.М.
 Статья на сайте eLibrary

В обзоре излагается подход к эмпирической реконструкции динамики сложных систем на основе прямого анализа временных рядов данных. Методы рассматриваются на примерах климатических процессов и позволяют выявлять механизмы наблюдаемой динамики, оценивать влияние внешних факторов и строить прогнозы будущего поведения системы.

На английском языке
Data-driven modeling of nonlinear phenomena in climate
Feigin A.M., Mukhin D.N. and Loskutov E.M.

We study an approach to the data-driven reconstruction of complex system dynamics being developed at the Institute of Applied Physics of the Russian Academy of Sciences, which allows, based on direct analysis of time series data, identifying both the mechanisms underlying the observed dynamics and predicting their changes under the influence of various factors. Using examples of the analysis of various climatic processes (such as El Ni\ no phenomenon, climate variability during the Pleistocene, the dynamics of the earth's atmosphere), methods for solving problems that arise at different stages of empirical reconstruction are discussed: data preprocessing and selection of dynamic variables, constructing an evolution operator, studying the nonlinear dynamic properties of the system and analyzing possible scenarios for its future behavior.

DOI: https://doi.org/10.52452/00213462_2026_69_09_770