Providing Classification Map of PM۲.۵ Air Pollution Using Decision Tree and Random Forest methods

Hazardous chemicals escape to the environment by a number of anthropogenic activities and may cause adverse effects on human health and the environment. Air pollution is caused by pollutants, such as carbon monoxide, Sulphur dioxide, nitrogen oxides, volatile organic compounds, ozone, heavy metals, and respirable particulate matter (PM۲.۵ and PM۱۰). An important part of air pollution is related to the pollution of suspended particles less than ۲.۵ microns (PM۲.۵). In order to control and reduce air pollution, classification of PM۲.۵ is of great importance to achieve sustainable development in cities. The influence of external factors on the spread of air pollution is not hidden from anyone. External factors that are used for classification pollution in this paper were meteorological parameters (includeding wind speed, wind direction, temperature, relative humidity, air pressure, and rainfall), topographic status, intensity of temperature inversion, pollution for the two nearest stations, and the time dependence. Among the classification methods, tree-based methods were chosen due to perception and simplicity. In this paper, CART and C۴.۵ decision tree algorithms, as well as random forest algorithm were used to classify the PM۲.۵. Among these methods, random forest method with the overall accuracy of ۷۸.۹% and the kappa index of ۷۵.۱% was selected as the best classification method. The results showed that the pollution parameters of the two nearest neighbors, topography, temperature, air pressure, rainfall, intensity of temperature inversion, relative humidity, wind speed, wind direction, month of the year, day of the week, hour of the day have the greatest impact on the classification of the superior method, respectively.

Keywords:

Air pollution , Decision Tree , Random Forest , PM۲.۵ , Overall accuracy , Kappa index.