Analysis Of Air Quality Data Based On Complex Network Theory

With the rapid development of the global economy,the air quality has been seriously damaged.PM2.5 is an important indicator of air quality.It is also the main pollutant in air pollution,and has received wide attention.As the capital of China,Beijing has a typical problem of urban air pollution.It economy develops rapidly.This paper selects the air quality data of Beijing as the research object.Based on the theory of the mutual conversion between time series and complex networks,the time series of Beijing air quality is mapped into complex networks.Topological characteristics of complex networks are studied to further explore the intrinsic characteristics of time series.By analyzing the topological characteristics of complex networks and the physical and geographical characteristics of the monitoring sites in Beijing,the main factors affecting air quality are analyzed.Contents of this paper mainly includes the following two aspects.(1)Time series analysis of air quality based on correlation coefficient network.Taking the air quality data of eight monitoring stations in Beijing as the research sample,the air quality time series is reconstructed by phase space reconstruction.Then the phase points in the phase space.The optimal delay time and embedding dimension in the phase space are determined by the C-C algorithm.By using the correlation coefficient method,we determine the connection between the phase points and the edges of the complex network by setting critical threshold values.The air quality time series is mapped to an undirected and unweighed complex network.Based on the analysis and comparison of the three topological characteristics(degree distribution,average clustering coefficient and modularity)of the complex network,the optimal critical threshold values is obtained.Finally,we determine the optimal correlation coefficient network of air quality.Based on the topological properties of the optimal complex network,we use the K-means clustering to obtain three kinds of clusters.For monitoring points in each cluster,combined with their natural geographical features,the main factors affecting air quality are analyzed.(2)The visibility graph of air quality.Based on the visibility graph algorithm,we map the air quality time series of eight monitoring stations in Beijing to the complex network.The topological characteristics of the eight complex networks are obtain.We find that the topological characteristics of the visibility graph network are different from those of the correlation network.The air quality time series of Huairou is mapped to undirected complex networks,using visibility algorithm.Based on the visibility graph networks,we analyze the reasons for the formation of large nodes and reveal the community structure of complex networks.By exploring the small world characteristics of complex networks,we find that all of the five mapped complex networks are of small-world,and the corresponding mapped network of 2013 presents the most obvious characteristics of small world.