MODELING RESULTS OF THE MIDDLE-TERM FORECAST OF THE AIR QUALITY INDEX IN BISHKEK

Authors

  • N. M. Lychenko Machinery researching and Аutomatics Institute of Kyrgyz Republic National Academy of Science
  • L.I. Velikanova Machinery researching and Аutomatics Institute of Kyrgyz Republic National Academy of Science

Keywords:

artificial neural network FFBNN, training sample, air quality index, meteorological factors, polluting factor, forecasting horizon, forecasting error.

Abstract

A methodology has been developed for forecasting the AQI air quality index in Bishkek for 24 hours in advance, taking into account meteorological factors and the polluting factor (the number of tons of coal burned daily at the city's CHP). The results of modeling on the basis of a artificial feedforward neural network using the backpropagation algorithm (FFBNN) for 3-hour AQI forecast periods of summer and winter periods for forecasting horizons up to 24 hours are presented. It is shown that the accuracy of forecasting 24 hours ahead for all forecast periods compared to the short-term forecast (3 hours ahead) deteriorated by 34% for the summer period and 44% for the winter period.

References

Air Quality Index (AQI) - A Guide to Air Quality and Your Health. US EPA. 9 December 2011.

Великанова, Л. И. Мультирегрессионные и обобщенно- регрессионные нейросетевые модели краткосрочного прогноза загрязнения РМ2.5 в г. Бишкек с учетом метеорологических параметров / Л. И. Великанова, Н. М. Лыченко // Проблемы автоматики и управления. – 2019. – № 2(37). – С. 16-25. – DOI 10.5281/zenodo.3594777. – EDN MNNDYU.

Великанова, Л. И. Моделирование краткосрочного прогноза загрязненности воздуха твердыми частицами на основе искусственных нейронных сетей с учетом фактора загрязнения / Л. И. Великанова, Н. М. Лыченко // Проблемы автоматики и управления. – 2022. – № 3(45). – С. 110-118. – EDN TGNBLC.

Donald F.Specht A general regression neural network //IEEE Transactions on neural networks. — November 1991. — Vol. 2, № 6. 568–576. DOI:10.1109/72.97934

Widi Aribowo. Optimizing Feed Forward Backpropagation Neural Network Based on Teaching-Learning-Based Optimization Algorithm for Long-Term Electricity Forecasting // International Journal of Intelligent Engineering and Systems, Vol.15, No.1, 2022. DOI: 10.22266/ijies2022.0228.02

Верзунов, С. Н. Применение глубоких нейронных сетей для краткосрочного прогноза дальности видимости / С. Н. Верзунов // Проблемы автоматики и управления. – 2019. – № 1(36). – С. 118-130. – DOI 10.5281/zenodo.3253019. – EDN IFNJHG.

Karimian H, Li Q, Wu C, et al. Evaluation of different machine learning approaches to forecasting PM2. 5 mass concentrations // Aerosol and Air Quality Research.- 2019.- 19(6). –p. 1400-1410. DOI: 10.4209/aaqr.2018.12.0450

Tong W, Li L, Zhou X, et al. Deep learning PM2. 5 concentrations with bidirectional LSTM RNN // Air Quality, Atmosphere & Health.- 2019. - 12(4). –p. 411-423. DOI: 10.1007/s11869-018- 303 0647-4.

AirNow Depatment of State // https://airnow.gov/index.cfm?action=airnow.global _summary #U.S._Department_of_State $Bishkek (дата обращения: 05.10.2022)

Cайт «Расписание погоды rp5.ru» Архив погоды в Бишкеке https://rp5.ru/%D0%90%D1%80%D1%85%D0%B8%D0%B2_%D0%BF%D0%BE%D0%B3%D0%BE%D0%B4%D1%8B_%D0%B2_%D0%91%D0%B8%D1%88%D0%BA%D0%B5%D0%BA%D0%B5 (дата обращения: 05.10.2022)

Современное прогнозирование. URL: https://forecasting.svetunkov.ru/etextbook/ (дата обращения: 30.09.2022)

Downloads

Published

2023-05-04

Issue

Section

INFORMATION TECHNOLOGY AND INFORMATION PROCESSING

Categories