Research On Data Mining Algorithms For Air Pollution Source Screening

Air pollution is one of the major threats to human health.How to control air pollution has also become the focus of attention and research by the government,the public and scholars.Space grid monitoring provides support to some extent to understand the causes of air pollution and to effectively manage it by omni-directional air related component statistics.At the same time,the potential value of monitoring data has not been effectively exploited.Therefore,this paper studies the data mining algorithm for air pollution sources under the condition of systematic analysis of air grid monitoring data from the perspective of multidisciplinary and multi-disciplinary integration of statistical learning,machine learning,and decision theory.The main research contents are as follows:(1)Research on air pollutant contribution evaluation model based on Multiple Attribute Decision Making:In view of the limitations of existing air pollutant contribution evaluation methods that can not effectively quantify the contribution of various air pollutants under different pollution sources,an improved CRITIC(Criteria Importance Through Inter-criteria Correlation)model for air pollutant contribution evaluation algorithm is proposed.Firstly,under the condition of systematic analysis of air grid monitoring data,the deviation index in the maximum deviation method is used as the measurement index of information in the CRITIC method to determine the weight of six air pollutants.Then,the comprehensive evaluation indicators are sorted by the Simple Additive Weighting Method(SAW)method.Thereby,an air pollutant contribution evaluation model based on multi-attribute decision making is constructed.Finally,the feasibility and effectiveness of the algorithm are verified by combining dust cases in a certain area.(2)Research on air pollution source category mining based on Mountain Method algorithm:Because of the problem of low clustering performance caused by random selection of initial clustering center in existing air pollution source category mining,a mining model of air pollution source category based on Mountain Method algorithm is proposed.Firstly,under the condition of analyzing air pollution data,Principal Components Analysis(PCA)is used to identify the main air pollution components.Then,the Mountain Method algorithm is applied to find k initial clustering centers of air pollution components.Thirdly the k initial clustering centers and K-Means algorithm are utilized for cluster analysis.Thus we can obtain the main categories of air pollution sources.Finally,a comparison experiment is conducted between the traditional mining model of air pollution source categories and the mining model of air pollution source categories based on Mountain Method algorithm to verify the validity of the algorithm.