Research On Air Quality Prediction Method Based On Air Pollutant Propagation Path Analysis

Air quality is affected by many factors and shows a complex change pattern.The changeable air reaction system makes the accurate prediction of air pollutant concentration face severe challenges.For the prediction of hourly air pollutant concentration,considering the spatial-temporal dynamic relationship of air pollution,analyzing the source of air pollutants and the transmission law of different transport tracks will be very useful for the accurate prediction of regional air quality.Based on this,this paper proposes an air quality prediction method based on the analysis of air pollutant transport path,which fully considers the spatial-temporal dynamic relationship of air quality evolution characteristics,and aims to build a more accurate and stable air quality prediction model to reveal the evolution trend of air quality.First of all,this paper analyzes the spatial and temporal dimensions of air quality characteristics to explore the main factors affecting air quality.For the candidate input data set,the machine learning system based on the lifting tree is used to select the features of the original data,and the importance ranking of the influence degree of the regional features on the air pollutant concentration is obtained,and the important input features affecting the air pollutant concentration are screened and determined.Secondly,according to the characteristics of air pollutants in different places,using the air quality ground monitoring data and the data of global data assimilation system,the paper studies the sources of imported air pollutants in Beijing and the key transport path based on the backward air flow trajectory analysis method,and excavates the spatial laws of air pollutants in different places.This Law provides a strong support for the selection of input characteristics of air quality prediction model.Thirdly,considering the results of pollutant source and transport path analysis,taking into account the temporal and spatial attributes of air pollutant concentration sequence,a Long Short-Term Memory neural network embedded attention mechanism is constructed to reasonably represent the relationship between meteorological characteristics and air pollutant concentration,and finally predict the air quality of the region.In this paper,the rationality of the proposed model is verified by constructing a sequence prediction problem for one or more time series in the future.Finally,the method proposed in this paper is verified.The experimental data used the real meteorological data and air pollutant concentration data from January 1,2017 to June 30,2019 in Beijing Tianjin Hebei region,verified the law of air pollutant transport in different places,and evaluated the effect of the prediction model from multiple perspectives.The experimental results show that the method proposed in this paper has certain rationality and good reference value.