Chapter 4 - Intelligent systems in air pollution research: a review

Abstract

Air pollution is one of the most serious challenges human society currently faces. It is not only harmful to humans, but also has negative impacts on animals, plants, and structures like buildings. To deal with air pollution, predicting the amount of contaminants in the air plays a vital role. Therefore this chapter provides a brief introduction to air pollution, and the use of artificial intelligence and machine learning approaches to forecast the amount of air pollution. Initially, the main concepts are given, and then, a quick literature review is presented. Next, an artificial neural network is employed by the research team to predict the amount of NO2 and PM10 for Cologne, Germany, as a case study. For the considered case study, a comprehensive error analysis is carried out by taking different error-related indicators, including absolute error, root mean square error, and their normal values, in addition to the coefficient of determination. Moreover, to give a much better insight, an error analysis is carried out for the smaller range of effective input parameters, in addition to the whole range. According to the results, the coefficients of determination for forecasting NO2 and PM10 by the developed models are 0.904 and 0.911, respectively, which are up to 10% higher than for the developed models from other studies (with values of around 0.8).