Preliminary study on the method of particle concentration curve fitting by LOESS with composite robust weight

Particulate matter is one of the main sources of air pollution, and the monitoring of its concentration has direct significance for the judgment of air quality. By research the numerical fitting method of particulate mass concentration, this paper chooses the LOESS method combined with robust weights to carry out numerical fitting, which not only retains the trend of the original monitoring value, but also eliminates the significant abnormal data. Through several iterations of fitting, smooth particle concentration curve is finally generated, which provides a visual basis for research particle concentration changes.