The Research On Air Quality Prediction And Assessment Based On Uncertainty Theory

As socialism with Chinese characteristics enters a new era,the Party Central Committee regards the construction of ecological civilization as the fundamental plan related to the sustainable development of the Chinese nation.Air quality,as one of the core indicators of ecological civilization construction,is increasingly receiving attention and emphasis.Currently,various cities have built urban air quality monitoring networks based on air monitoring stations,enhancing the ability to detect air quality and providing data support for air quality research.Researchers usually directly use the average concentration of air pollutants from various monitoring stations as the urban overall air pollutant concentration for air quality research.However,this method has certain limitations.Since the urban overall concentration of air pollutants is not a definite value,but a variable within a certain range,the data collected by monitoring stations is a series of observations of the variable,and the average value obtained from the limited monitoring station data cannot accurately reflect the distribution of air pollutant in the entire city.In response to the above issues,the thesis introduces uncertainty theory to process the data collected by monitoring stations and conducts air quality prediction and evaluation research based on uncertain statistical methods.The main research content and innovations are as follows:(1)In view of the uncertainty of air quality related data,the concept of uncertain variables is introduced.The urban overall air pollutant concentration and AQI are described as uncertain variables,and the data collected by monitoring stations is a series of observations of the uncertain variables.By using the method of uncertainty moment estimation,the distribution of these uncertain variables is obtained and used to describe the overall air quality distribution of the city.(2)The thesis studies the problem of air quality prediction based on uncertainy theory.Firstly,the uncertain regression model is introduced,and an uncertain multivariate regression model of air quality in Beijing for the entire year of 2022 is established.Considering the significant seasonal characteristics of air quality in Beijing,uncertain multivariate regression models of air quality are established for each season.Subsequently,the uncertain time series model is introduced,and an uncertain autoregressive time series model of air quality in Beijing is established.The application of these two models in air quality prediction has achieved good prediction results.The actual observation values are all within the confidence interval estimated by the model,indicating that both methods are practical and reliable.(3)The thesis studies the problem of air quality evaluation based on uncertainy theory.Firstly,the uncertain k-means clustering model is proposed and established to perform clustering evaluation on the air quality data in Beijing for the year of 2022,dividing the air quality into five different levels and analyzing the air quality status based on the clustering results.Under the uncertain k-means clustering evaluation system,Beijing achieves a score of good or moderate air quality for 91.51 percent in the year of 2022.Subsequently,the uncertain k-nearest neighbor classification model is proposed and established to perform classification evaluation on air quality in Beijing,achieving an accuracy rate of 0.9128 and obtaining good classification evaluation results.