LSTM Forecasting: Time Series Forecasting to Predict Concentration of Air Pollutants (CO, SO2, NO and NO2) in Krasnoyarsk, Russia


Тип публикации: доклад, тезисы доклада, статья из сборника материалов конференций

Конференция: Computer Science Online Conference, CSOC 2021

Год издания: 2021

Идентификатор DOI: 10.1007/978-3-030-77448-6_17

Ключевые слова: air pollution, carbon monoxide, krasnoyarsk, lstm, nitrogen dioxide, prediction model, sulfur dioxide, time series forecasting

Аннотация: The national weather agency says Krasnoyarsk had the dirtiest air of any Russian city in 2018, beating out Mumbai and Guangzhou. Krasnoyarsk has air so bad that the authorities regularly warn people to remain inside, avoid walking and sporting outdoors. During periodic “black sky” events, Krasnoyarsk’s 1 million residents suffer frПоказать полностьюom toxic levels of smog in winter, when coal-powered emissions peak, and in late summer smoke from wildfires spread. An issue that is magnified due to a lack of effective air pollution prediction techniques, this study has used the raw numerical data of key pollutants to predict their future status through LSTM (Long Short-Term Memory) modeling. Machine learning methods trained on raw numerical time series data exhibit fundamental limitations such as a high sensitivity to the hyper parameters and even to the initialization of random weights. The article shows potential of deep learning models with high accuracy using the time series data of air pollutants and meteorological parameters #CSOC1120. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Ссылки на полный текст


Журнал: Lecture Notes in Networks and Systems

Выпуск журнала: Vol. 228

Номера страниц: 191-198

ISSN журнала: 23673370

Издатель: Springer Science and Business Media Deutschland GmbH


  • Kulagina L.V. (Siberian Federal University (SibFU), Krasnoyarsk, Russian Federation)
  • Kulagina T.A. (Siberian Federal University (SibFU), Krasnoyarsk, Russian Federation)

Вхождение в базы данных