Establishments Of BP Neural Network Model Based On The Improved PSO With Applications To Air Quality Prediction

Daily monitoring of air quality,accurate data analysis,quality prediction and visualization are conducive to the comprehensive control of urban air quality.The aim of this paper is to explore the method of intelligent air quality evaluation,improve the algorithm of intelligent air quality evaluation and establish the model of intelligent air quality evaluation.Firstly,the basic principle of air quality assessment,the air quality classification corresponding to the air quality index and the information contained in each characteristic parameter are described.Secondly,the basic model of BP neural network,the selection in the past January to February,April to May of 2020,and January to February of 2021 in Chongqing NO₂、CO、O₃、SO₂、PM2.5、PM10 and AQI concentration information as sample set,and the data were normalized processing,in order to eliminate the influence of dimensional and orders of magnitude,and then correlation analysis to eliminate associated with air quality index AQI smaller indicators,in order to simplify the input data,lays the foundation for the intelligent evaluation model is set up.Secondly,in order to take advantage of the global optimization characteristics of particle swarm optimization algorithm,a new model is proposed to improve it.The new model is based on the K-means clustering algorithm to improve the diversity of the population.Taking several optimization test functions as the research object,the optimization simulation results show that the improved PSO algorithm has stronger global optimization ability and can avoid the particle falling into the local extreme value to a greater extent.Finally,the particle dimensions in the improved PSO are consistent with the number of weights and thresholds in the BP neural network,the mapping relationship between improved PSO search space and neural network weights and thresholds is established.After the topology of the neural network is determined,the improved PSO can be used for training and testing.In order to verify the effectiveness of the improved algorithm proposed in this paper,a comparative simulation experiment is carried out with the traditional BP neural network algorithm.The results show that the accuracy and operating efficiency of the improved algorithm are improved to varying degrees,which verifies the superiority of the improved algorithm.Secondly,the network simulation test and the actual air quality index are compared and verified.The results show that the state information displayed by the simulation is basically consistent with the actual air quality,and the accuracy rate is more than 85%.This shows that the model proposed in this paper is practical and can be used to evaluate the air quality in Chongqing,which is helpful to the development of environmental monitoring.