A neural network
approach to forecast particulate matter concentration in Manali area of Chennai
City
Nadeem Imran and Sheik Uduman P.S.
Res. J. Chem. Environ; Vol. 27(8); 35-42;
doi: https://doi.org/10.25303/2708rjce035042; (2023)
Abstract
Air pollution is one of the threatening menaces confronting all over the globe in
recent decades. Among the air pollutants, PM2.5 is one of the major alarming components
rising at a rapid pace in the metropolitan city of Chennai due to the fast expansion
of industrialization and urbanization. An early observation for tackling the rise
in particulate matter concentration (PM2.5) levels requires precise prediction.
In this regard, the present study employs a Multilayer Perception Neural Network
(MLPNN) technique using the Levenberg-Marquardt optimisation training algorithm
for forecasting the one-day concentration PM2.5. The model evaluation statistics
R2 and index of the agreement have been utilized for assessing the forecasting accuracy.
The results confirm that the ANN6 model is the most suitable for acquiring the almost
error-free model to achieve real-time forecasting of PM2.5 concentration.
