Prediction of air pollution with machine learning methods

Air pollution is accepted as a worldwide risk to humans. Prolonged exposure to high levels of ozone pollutants can lead to chronic respiratory diseases such as bronchitis, emphysema, asthma, etc. In addition to its effect on the human body, high levels of ozone also affect the photosynthetic efficiency of crops, resulting in reduced crop yields. In addition, it is recognized as one of the key pollutants that degrade air quality in urban areas. Therefore, predicting air quality previously plays an important role in warning and controlling peoples about air pollution. In this study, hourly ozone air pollutant concentration values in Bursa Uludag University and Kulturpark stations for Bursa province were estimated by machine learning algorithms. The data were obtained from the National air quality monitoring network site of the Ministry of Environment, Urbanization and Climate Change. Pollutant and meteorological data (air temperature, wind speed, relative humidity and air pressure) were used in forecasting model. Random forest, decision tree, support vector, k-nearest neighbor and multilayer perceptron regression were used as the machine learning methods to forecast the ozone values. The root-mean-square error, mean squared error, mean absolute error, mean absolute percentage error, and coefficient of determination were used to evaluate the performance of the regression models. It was seen that the random forest regression algorithm for two stations gave better results in estimating ozone concentrations than other algorithms.