Analysis Of Influencing Factors And Forecast Of Air Quality Index In Wuhan

In recent years,China's industry has embarked on a high-speed development path and achieved rich results,but the environmental governance issues have been ignored during the development process.A large number of industrial emissions and automobile exhaust have adversely affected China's air quality,leading to Many cities have severe air pollution.Some tiny particles in the haze may be inhaled by residents,and then chemical reactions occur in the heart,lungs,and throat,causing damage to the human immune system and organs,and may also cause allergies and lung diseases.The haze not only contains a lot of harmful particles,but also greatly reduces the visibility of the outdoors,which has a serious impact on residents' transportation.Therefore,air treatment is urgent.In 2017,Premier Li Keqiang raised a blue sky defense battle,and air pollution began to become a matter of close concern in China.For the government,in order to manage the air pollution problem,the monitoring and forecasting of air quality is extremely important.It is of great significance to predict the air quality trend of a city and help the country to protect and develop the ecological environment..There are many influencing factors and complex causes of air pollution.The degree of air pollution in our daily life depends on many factors,such as the concentration of various pollutants in the air,different wind directions and weather conditions in a day.Taking Wuhan as an example,this paper selects hourly data of air quality index,air pollutant concentration,and meteorological factors from January 1,2017 to December 10,2019,and performs missing value interpolation and outlier detection on the data.Pre-processing,data standardization,etc.,then find the significant factors that affect the air quality index through linear regression and correlation coefficients,and select significant features for the next multivariate prediction.In order to better predict the air quality index and improve the accuracy of the air quality index,this paper builds a combination model based on LSTM neural network,STL decomposition,and elastic network regularization to predict Wuhan air quality index.In the LSTM combination model,the STL decomposition is first used to decompose the Wuhan Air Quality Index into a trend term,a season term,and a residual term.Then,each component is predicted using an LSTM model with a self-attention mechanism.The elastic network regularization is added to reduce the dimensionality of the features.Finally,the prediction results of each component are integrated to obtain the final prediction result.In order to verify the performance of the combined model proposed in this paper,an econometric model and a machine learning model are introduced for comparison.The experimental results show that the LSTM combination model can accurately predict the trend of air quality in Wuhan,and is superior to traditional econometric models and machine learning models in accuracy and robustness.It can be seen that the introduction of STL decomposition can greatly improve the fitting accuracy of the model.At the same time,adding elastic network regularization to the multivariate time series prediction can improve the generalization and prediction accuracy of the model,which has great practical significance and feasibility.