Developing Soft Computing based Models for Prediction of Pollutant PM10 of Air

The three major pollution causing components of environmental air are SO
₂
concentration, NO
₂
concentration and the particulate matter (PM). The particulate matter is comprised of solid particles mixed with drops of liquid. PM
₁₀
contains particles having diameter equal to or less than 10 micrometer, and PM
₂
.
₅
contains such particles having diameter smaller than or equal to 2.5 micrometer. These particles are extremely dangerous to human health. It would be very interesting if the prediction of the value of PM
₁₀
is done on the basis of present SO
₂
and NO
₂
concentrations in air. This has been a motivating factor to write this research paper, wherein an attempt has been done to design soft computing based prediction models, using MATLAB, for PM
₁₀
of the air. The fuzzy and ANN (artificial neural network) based models have been trained using an available air quality secondary data, of year 2019 (When the air pollution level was at its peak), of some major cities of Uttarakhand, India. Finally the performances of the developed prediction models have been compared.