Research On The Application Of Neural Network In Urban Air Quality Prediction

According to the relevant data of China's Ministry of Ecological Environment,the compliance rate of urban ambient air quality in 2020 is 59.9%,but the improvement of air quality is not stable.Regional severe pollution weather such as acid rain and haze still occurs frequently,which seriously affects people's daily life and health.The prediction of air quality can provide valuable reference for government management and people's life,which is of great significance to guide people's daily life.Based on the air quality monitoring data of Huzhou City from 2014 to 2021,this paper carried out relevant research on air quality prediction.The main research contents are as follows :(1)PCA-MEA-BP model: To improve the slow convergence speed of the BP neural network and easy to fall into local optimum,this paper uses Principal Component Analysis(PCA)to reduce the dimension of data,and uses the Mind Evolution Algorithm(MEA)to optimize the weight and bias of BP neural network.It can be obtained from the comparative experiments that the PCA-MEA-BP model has the best prediction effect on the Air Quality Index(AQI)of Huzhou City,which verifies the effectiveness and superiority of the PCA-MEA-BP model.(2)LPP-BP model: Aiming at the problem of slow convergence speed of BP neural network,this paper uses Local Preserving Projection(LPP)to reduce the dimension of data,which can retain more local structure information between data.To improve the prediction effect of the model,this paper also proposes a method to predict AQI in spring,summer,autumn,and winter.The experimental results show that the LPP-BP model based on seasonal prediction has higher prediction accuracy and practical significance.(3)BP-LSTM combination model: Aiming at the problem that BP neural network can not take into account the timing information of AQI,this paper combines the BP neural network and LSTM neural network by weight to realize the complementary advantages of the model.The comparative experiment between the BP-LSTM combination model and the single model is designed to verify the effectiveness of the BP-LSTM combination model.(4)Air Quality Index Prediction System: An air quality index prediction system is developed based on BP-LSTM combination model.The system uses B/S architecture,carries out system planning and design from data layer,business logic layer and data representation layer,and provides city AQI query,city history AQI query through data visualization method,and shows the future trend of city AQI.