Research and application of an evolutionary deep learning model based on improved grey wolf optimization algorithm and DBN-ELM for AQI prediction

Accurate forecast of air quality index (AQI) can provide reliable guarantee for air quality early warning and safe production. In this paper, a hybrid model for predicting AQI is presented. Firstly, the original AQI data is decomposed into multiple intrinsic mode functions (IMFs) components by using time varying filter based empirical mode decomposition (TVFEMD). To reduce the amount of calculation, sample entropy (SE) is introduced to estimate multiple IMF components. Secondly, the grey wolf optimization (GWO) algorithm was improved, the dimension learning-based hunting (DLH) search strategy was introduced to avoid falling into local optimum. Meanwhile, the opposite search strategy was introduced in the initialization of DLH strategy to enrich population information. Thirdly, the parameters of deep belief network - extreme learning machine (DBN-ELM) model is optimized by IGWO algorithm. Then the DBN-ELM model with optimal parameters are used to forecast each IMF component, respectively. Finally, the predicted value of each IMF component is reconstructed to get the total AQI predicted value. The comparison between the presented model and the other benchmark models used in this paper shows that presented model is better than other model in accuracy and generalization, which demonstrates that the presented model can effectively predict AQI.