Artificial Neural Network and Hybrid ANN Models for the Prediction of Particulate Matter (PM10) Concentration, Trivandrum (India)”,

Air pollution rate is getting verse day by day around the world due to industrialization and urbanization. And particulate matter (PM) is considered as one of the major contributor to this increase in air pollution. Besides the deposition of trace elements in air and reduction in visibility, the direct impact of particulate matter on vegetation and human health are serious issues. In several researches a reliable relation was found between health effects and elevated concentrations of atmospheric PM10 concentration. So for the risk-impact assessments and health studies it is very important to quantify the air pollutant concentration rate in the atmosphere and forecast particulate matter concentrations. In the present study particulate matter concentration of PM10 (10 micron size) was predicted using Artificial Neural Network and a Hybrid Artificial Neural Network. The parameters which affect particulate matter concentration such as Temperature maximum, Temperature minimum, Rainfall, Relative humidity and Station level pressure were considered as the input parameters in the modeling. ANN and hybrid ANN models were developed for Trivandrum district and the model performance was compared by statistical evaluation.