Air Quality Index Prediction Based On The Hybrid Model SD-DEWOA-SVR

In this paper,a hybrid model SD-DEWOA-SVR based on the secondary decomposition-integration is proposed,which combines variational modal decomposition(VMD),complementary ensemble empirical mode decomposition(CEEMD),and sample entropy(SE),differential evolution whale optimization algorithm(DEWOA)and support vector regression(SVR)to predict the air quality index.Firstly,the data are decomposed by the secondary-decomposition(SD)algorithm which combines VMD and CEEMD algorithm,and obtain several eigenmodes.Secondly,SVR is selected as the prediction method in this paper,and the parameters c and g in SVR are optimized by DEWOA.Finally,the prediction results of all modal sequences are integrated to obtain the final predicted value of the air quality index.In order to better illustrate that the hybrid model SD-DEWOA-SVR proposed in this article has good stability and generalization ability,this article uses the air quality index(AQI)data of two sets of cities(Shanghai and Changchun),and compares the models with the other seven sets to verify.Results show that the SD-DEWOA-SVR model has better performance than other comparison models on the air quality index(AQI).