evaluation and forecasting of pm10 air pollution in chennai district using wavelets, arima, and neural networks algorithms

The advent of advanced features of soft computing can be used to solve complex problems which are more non-linear and messy. many of the applications have been analysed and validated by the researchers through soft computing approach in the past.neural networks (nn) with appropriate selection of training parameters is implemented apart from conventional mathematical model. in this paper, analysis is made on the estimation of pm10 air quality in selected regions of chennai district by wavelet approach with energy spectrograms. after analysing the results, nn of multilayer feed forward back propagation algorithm forecasts the air quality of selected regions. discrepancies in selecting the training parameters of nn’s have been overcome by trial and error basis. this work will be helpful in proving the powerful tool of nn to forecast short term nonlinear parameters and the predicted results will give us the clear design of existing problem and thecontrol measures need to be implemented.