An improved method for identifying particulate matter pollution type based on decision tree in Wuhai
Years of monitoring data show that PM10 is the primary pollutant affecting the environmen-tal air quality in Wuhai City,and particulate matter pollution is the focus of air control.Identifying the air pollution type is the vital process to find pollution cause and conduct dynamic source appor-tionment.In this study,after determining pollution type and identifying criteria based on source ap-portionment result and pollutant emission data,with the air quality monitoring data of Wuhai,the types of particulate pollution in Wuhai including sand dust,fugitive dust,industry,secondary genera-tion and vehicle was accurately identified by multi-parameter feature ratio method.The CART deci-sion tree was used to improve the identifying process in order to overcome the subjectivity and com-plexity of pollution identifying threshold adjustment.After rapid improvement of pollution type identif-ying threshold based on decision tree,the accuracy of particulate pollution identifying was greatly im-proved.In 2021,the particulate pollution in Wuhai was mainly sand dust type,contributing 25.09%to PM10 and 15.71%to PM2.5.With this method,all the particulate standard exceeding days were identified and the accuracy rate of sand and dust weather identification could reach 81.82%.Based on de-cision tree,the selection of threshold value by multi-parameter feature ratio method can be effectively improved,and the type of particulate pollution can be more quickly and accurately identified.
Wuhaiparticulate pollutionpollution typedecision tree
万方数据
