AQI Trend Prediction Based On CEEMD And Error Correction Model

With the rapid development of economy,air pollution occurs frequently.It not only has a negative impact on human health and ecosystem,but also is not conducive to the sustainable development of economy.Currently,the prevention and control of air pollution is an urgent need in China,and air pollution prediction can provide reliable information for this process.Therefore,it is necessary to establish a reliable air pollution prediction model.And the widely used air pollution prediction model is the error correction model.However,this traditional error correction model still has room for improvement.Based on this background,this article proposes an improved error correction model to predict AQI.On the basis of the traditional error correction model,CEEMD(Complementary Ensemble Empirical Mode Decomposition),PSO(Particle Swarm Optimization),GSA(Gravity Search Algorithm),GWO(Grey Wolf Optimization),SVR(Support Vector Regression),and GRNN(Generalized Regression Neural Network)are used to establish an improved error correction model,which named CEEMD-SLM-ECM(Complementary Ensemble Empirical Mode Decomposition-Statistical Learning Model-Error correction model).In this paper the AQI data of 17 port cities in the 21 st Century Maritime Silk Road Economic Belt are selected to test the predictive ability of this model.Data analysis shows that the accuracy of the CEEMD-SLM-ECM model is better than the traditional error correction model.Taking Shanghai as an example,the MAPE range of the CEEMD-SLM-ECM model is 5.55%-10.83%,while the MAPE range of the traditional error correction model is 8.69%-28.44%.Therefore,the CEEMD-SLM-ECM model established in this paper is a reliable and accurate trend prediction model of air pollution.