Анализ и прогнозирование временных рядов: методы и приложения

2026/2027

ID 1121358

Статус: Маго-лего

Кто читает: Департамент анализа данных и искусственного интеллекта

Где читается: Факультет компьютерных наук

Охват аудитории: для своего кампуса

Преподаватели: Шляхов Фёдор Сергеевич

Язык: русский

Кредиты: 6

Контактные часы: 54

Дополнительные материалы в LMS Задать вопрос

Аннотация

The course is dedicated to modern methods of time series analysis and forecasting. Students will get familiar with both classical approaches and advanced topics: machine learning, analysis of chaotic systems, and fractal analysis. Particular attention is paid to the analysis and forecasting of chaotic time series.

Цель освоения дисциплины

To form a systematic understanding and practical skills in students for working with time series of various natures, including complex chaotic and fractal times series, from the stage of preprocessing to building and verifying predictive models.

Планируемые результаты обучения

Decompose a series into components, check for stationarity, build and interpret ACF/PACF
Apply filtering methods to highlight trend and suppress noise. Justify the choice of filtering method
Select parameters and build ETS, ARIMA, GARCH models. Evaluate forecast quality.
Design, train, and evaluate models for forecasting and classifying time series
Distinguish chaotic and regular time series. Analyze and forecast chaotic time series.
Conduct R/S analysis and DFA, calculate and interpret the Hurst exponent
Possess methods for preprocessing and filtering time series
Build and interpret classical forecast models
Apply machine learning methods for forecasting and classifying time series
Conduct analysis and forecast chaotic time series using nonlinear dynamics methods
Conduct fractal analysis of a time series