Study On The Division Method Of Regional Air Quality Distribution Based On Spectral Cluster Data Mining

The study of air quality is a complicated system engineering.The propagation and interaction of regional air quality lead to the dynamics and complexity of the evolution process of air quality.Therefore,effectively characterizing the structure and spatial distribution of air quality,reflecting the internal correlation and overall effect of air quality,and revealing the evolution law of air quality have become the key to effectively analyzing air quality and preventing and controlling pollution in China.Based on the analysis of regional air quality propagation mechanism,a complex network model of air quality was established in this paper,and the spatial and temporal distribution and interactive influence of air quality were systematically studied by using spectral clustering data mining method.The main research contents are as follows:Firstly,the influence factors of air quality transmission are analyzed,including meteorological factors(such as temperature,weather,pollutant concentration,etc.)and geographical factors(such as latitude and longitude).Bartlett test and KMO method are adopted to analyze the correlation of air quality influence factors.Second,applying complex network theory to dynamic analysis of air quality,the local air quality abstracted into network node,local air quality interactive relationship between the abstract into edge,based on regional air quality structure characteristics of the complex network mapping method,the air quality of space-time distribution and interaction relationship,considering the change of time cumulative effect,the dynamic characterization of regional air quality distribution state.Thirdly,a new method of air quality region division based on spectral clustering is proposed.On the basis of the complex network,spectral clustering analysis was carried out on the nodes,and the correlation matrix of air quality generated by the complex air quality network was used as the connection matrix to improve the spectral clustering,so as to realize the clustering analysis of regional air quality characteristics.Finally,the air quality complex network model and the air quality region division method based on spectral clustering were tested and analyzed to verify the validity and accuracy of the analysis results,providing theoretical basis for air quality prediction and pollution prevention and control.