Prediction A Of Air Pollutants In Henan Province Based On Neural Network

The increasingly serious air pollution problem is one of the most urgent environmental problems to be solved in China.Based on neural network model and particle swarm optimization algorithm in the study of internal mechanism,according to the sample data in 17 cities of Henan province,aimed at deficiencies in low prediction accuracy of neural network prediction model and particle swarm optimization algorithm and error existing large,slow convergence problems,based on improved particle swarm optimization algorithm of neural network prediction model.The main research work of this paper is as follows:Firstly,the internal structure of the neural network can be adjusted through programming basing on the supervised forward neural network model,According to the different input variables and activation functions in the hidden layer of the neural network,four different neural network model structures are established.Secondly,the internal mechanism of particle swarm optimization algorithm is studied basing on programming.The Levenberg-Marquardt algorithm was coupled with the particle swarm optimization algorithm.In Matlab,the optimized parameters in the neural network obtained by the Levenberg-Marquardt algorithm were used as the initial value of the particle swarm optimization algorithm to propose the improved particle swarm optimization algorithm.The error function of particle swarm optimization algorithm can be selected by programming.Finally,respectively using Pearson correlation analysis and cluster analysis to data preprocessing in 17 cities in Henan province,find out the correlation is stronger,the categories of similar data,and then choose the model to be used in the data sample,the improved particle swarm optimization algorithm and standard particle swarm optimization algorithm in the application of comparative analysis,the results show that better performance of the former.On this basis,four kinds of neural network air pollution prediction models based on the improved particle swarm optimization algorithm are established,and the four established models are applied in Henan province,respectively,to analyze and compare their prediction results.The results show that model 2 has the best prediction effect,followed by model 3,and model 1 and model 4 have poor prediction effect.In this paper,an effective neural network air pollution prediction model based on improved particle swarm optimization algorithm is established in Henan province,which not only enriches the theory of neural network model and particle swarm optimization algorithm,but also provides a certain reference for air environment governance in Henan province.