A Novel Approach for Prediction of Air Pollutant Concentration

Air pollution is one of the biggest concerns India is facing today. It has been increased due to urbanisation and industrialisation. In this paper, the proposed model uses time series based forecasting for predicting air pollutant concentration. Studies have proved that using linear and nonlinear model together will greatly improve the performance. Therefore, a hybrid algorithm has been proposed to use in both the above-mentioned models. Results show that the approach of dividing the original data and merging these individual models play a key role in improving the performance. ARIMA (Auto Regressive Moving Average) and ANN (Artificial Neural Network) are considered here. This hybrid method have increased the forecasting accuracy in time series applications.