Air Quality Measurement Based On Combined Neural-Bayesian Network Model

With the comprehensive implementation and implementation of the "air pollution ten" policy,the overall air quality has been greatly improved,and the proportion of fine days in key cities has been increasing year by year.However,the task of improving air quality is still arduous,and we need to continue to defend the blue sky,and clarify the governance ideas and specific tasks.In terms of air quality monitoring,air quality index(AQI)has always been the focus of attention.Combined with LSTM neural network and BN bayesian network,the lstm-bn combination model is built for the evaluation and prediction of ambient air quality.Firstly,in determining the factors affecting the AQI,this paper selects six air pollutants,namely PM2.5,PM10,SO2,NO2,O3 and CO,as the impact factors,and adopts the data of changsha city from 2014 to 2018 for model training.In the process of model training,EMD algorithm of empirical mode decomposition was implemented to solve the nonlinear characteristics of air pollutant time series data,and the advantages of LSTM neural network in processing time series data were used to realize the effective prediction of various air pollutants,and its performance was all greater than that of the comparison model.Finally,the bayesian network model structure of air quality in changsha city was learned by mountain climbing algorithm,and the output data of LSTM was used as input data to evaluate and predict the air quality index in the whole year of 2019.It is concluded that the overall accuracy and quarterly accuracy based on the lstm-bn model are: 86.85%,87.78%,83.52%,86.96% and 89.13%,respectively,which are all higher than the bayesian network,single-factor index method and fuzzy comprehensive evaluation method used for comparison.This paper realizes the effective prediction and evaluation of air quality index of changsha city,and puts forward some specific Suggestions for the air quality status of changsha city in 2020,which has certain construction significance for the environmental management of changsha city.