Affiliations: [a] College of Big Data and Intelligent Engineering, Yangtze Normal University, Fuling, Chongqing, China
| [b] Chongqing University of Arts and Sciences, Yongchuan, Chongqing, China
| [c] School of Traffic and Transportation, Beijing Jiaotong University, Beijing, China
Correspondence:
[*]
Corresponding author. An Jianmei, Chongqing University of Arts and Sciences, Yongchuan, Chongqing, China. E-mail: ajmcqwu@sina.com.
Abstract: Atmospheric pollutant PM2.5 does serious harm to human health. It is one of important tasks in our country to reduce the pollution and protect people’s lives. For this purpose, accurate prediction of the pollution conditions is needed, and a model based on BP Neural Network Algorithm is proposed in this paper. By using the data of PM2.5 and meteorological parameters observed in Fuling, a mountainous suburban region in Chongqing, China from Jan.1, 2016 to Sep.1, 2017, the effects of temperature, humidity and wind speed on PM2.5 were first analyzed by the principal component analysis. Then a prediction model based on the BP neural network algorithm was built by using satellite remote sensing data, and the concentrations of the pollutants were explored by the model. The experimental results show that the standard deviation between the prediction results and average variance of the observed data is only 0.1218. Finally, the relationship between the number of hidden neurons and the absolute error is discussed. The prediction results show that the performance of BP neural network is better than that of regression model.
Keywords: Pulmonary particulate, air quality prediction, BP neural network
