Monitoring of PM(2.5) Concentrations by Learning from Multi-Weather Sensors

This paper aims to monitor the ambient level of particulate matter less than 2.5 [Formula: see text] m (PM [Formula: see text]) by learning from multi-weather sensors. Over the past decade, China has established a high-density network of automatic weather stations. In contrast, the number of PM monitors is much smaller than the number of weather stations. Since the haze process is closely related to the variation of meteorological parameters, it is possible and promising to calculate the concentration of PM [Formula: see text] by studying the data from weather sensors. Here, we use three machine learning methods, namely multivariate linear regression, multivariate nonlinear regression, and neural network, in order to monitor PM [Formula: see text] by exploring the data of multi-weather sensors. The results show that the multivariate linear regression method has the root mean square error (RMSE) of 24.6756 [Formula: see text] g/m [Formula: see text] with a correlation coefficient of 0.6281, by referring to the ground truth of PM [Formula: see text] time series data; and the multivariate nonlinear regression method has the RMSE of 24.9191 [Formula: see text] g/m [Formula: see text] with a correlation coefficient of 0.6184, while the neural network based method has the best performance, of which the RMSE of PM [Formula: see text] estimates is 15.6391 [Formula: see text] g/m [Formula: see text] with the correlation coefficient of 0.8701.