KRASNOYARSK, December 7. /TASS/. Scientists at Siberian Federal University have developed a method for predicting the quantitative composition of harmful substances in the air using neural networks, taking into account several types of meteorological data at once, the press service of the university told TASS on Monday.
"The authors have proposed a new way to predict the concentration in the atmosphere of compounds such as nitrogen dioxide, nitric oxide, sulfur dioxide and carbon monoxide. With prolonged exposure to the human body, they can deteriorate health and, ultimately, lead to chronic and malignant diseases. The use of neural networks (LSTM) combined with methods of mathematical modeling made it possible to predict with high accuracy the quantitative values of hazardous polluting components, as well as meteorological conditions," the university told TASS.
Predicting the composition of the air is a challenge for researchers as many factors have an influence on it: exhaust gases, industrial emissions, coal combustion and dust. Moreover, the speed and nature of the distribution of harmful substances in space are individual for each of them. The scientists in their study used "raw" numerical data on the main air pollutants, which were the result of periodic measurements of air quality in 2017-2019, made by atmospheric monitoring stations in Krasnoyarsk.
Having compared the concentration of substances with 10 types of meteorological data (temperature, humidity, wind speed, etc.), the authors developed the architecture of a mathematical model for training a neural network. As a result, it was possible to improve the accuracy of the forecast and automate the process of assessing the risks of increasing air pollution levels.
"The existing models for predicting air pollution without machine learning have serious drawbacks that do not allow them to be widely used for long-term predictions. Artificial intelligence technologies based on LSTM can process not only individual images, but also entire data sequences (speech, video, etc.), are able to store information for a given period of time to selectively change it. Due to a special "training" system, the created model is able to track most of the unexpected fluctuations in the level of air pollution", Lyudmila Kulagina, associate professor of the Department of Technosphere and Environmental Safety of Siberian Federal University, told TASS.
The scientists concluded that the use of different types of weather data can improve the accuracy of air quality predictions for other harmful compounds. The application of the new method will contribute to the development of effective ways to protect the environment and identify sources of pollution, the university explained to TASS.