Combination Model Based On ARIMA And BP Neural Network Research On Air Quality Forecast In Xi’an

With the rapid economic development,pollution sources and emissions have increased,making the problem of air pollution more prominent.The continuous deterioration of air quality has caused a huge impact on the sustainable development of the national economy and the health of the people.Predicting air quality can not only provide relevant functional departments with air quality index and pollutant concentration information for a period of time,helping it to be effective At the same time,it also promotes public awareness of the current severe forms of air quality and increases people’s enthusiasm for participating in air quality and even environmental protection.Therefore,the establishment of high-precision prediction models has an important role in the prevention and control of air pollution.This article aims to improve the prediction accuracy of air quality.Based on the existing research,it is found that the air quality index and the main pollutants have linear and non-linear composite characteristics.Taking the representative air quality index and the main pollutants as examples,the selection is made.The ARIMA model with linear prediction characteristics and the BP neural network model with non-linear prediction characteristics are used to predict a single model of air quality index and primary pollutants,and then the two are combined to construct an ARIMA-BP neural network combination model to predict it.In the combination model,two combination models with different combinations are proposed: series combination model and parallel combination model.Genetic algorithm is used to optimize the weight of the parallel combination model,and the prediction results of each model are compared.The conclusion shows that the introduction of genetic algorithm can effectively optimize the weight of the combined model,and the accuracy of the combined model after optimization is higher than that of the single model.It is confirmed that the air quality index and the main pollutants have linear and nonlinear composite characteristics,and the nonlinear characteristics are higher than linear feature.This paper constructs a combined forecasting model,hoping to provide a new reference method for accurate prediction of air quality.