Statistical Research On The Fluctuation And The Influence Factors Of Pm 2.5 In Three Cities

As the air pollution has become the most serious problem in many cities since October 2011,PM2.5 and Air Quality Index(AQI) get more and more concern among other pollution indices which can pose a great threat to public health and urban environment. These make the research of the inner variation rule and external influencing factors of PM2.5 is imperative and meaningful.This paper selects Beijing,Guangzhou, Chengdu as the research objects, using PM2.5 and AQI data to test the influence factors and fluctuation with wavelet analysis and RBF neural network.Through the analysis of three cities we can found that no matter which city, the pollutant density peaked during the night, and in the daytime, the concentration of PM2.5 declined gradually. Wavelet transform of PM2.5 can decomposited and restricted the original sequence, and it is obviously that the wavelet analysis can better predict the trend of fluctuations than the original sequence.In the analysis of influencing factors, RBF neural network model was used to analyze the influence of various factors on PM2.5, according to the related research, we found that the current research mainly focus on the chemical composition of PM2.5 through professional instrument to determine instead of using statistical analysis,but this paper select PM2.5 monthly data, electronic and oil industry output value and other indicators as the research object, the results showed that: passenger and freight volume is one of the biggest factors influence on PM2.5 fluctuations no matter which city.