Advances in urbanization and industrialization and increase in human activities have caused significant ecological and environmental effects in the recent past. Various prediction methods and techniques have been for early detection and reduction of air pollution. In this study, air quality data from Sichuan Province, China were collected from March 2015 to March 2020. EWM algorithm was used to determine the weights of factors that affect air quality such as PM2.5, PM10, SO2, CO, NO2, O3, monthly average precipitation, and relative humidity. EWM-BP, EWM-RNN, EWM-GRU and EWM-LSTM air quality information entropy prediction models were constructed based on the data from Sichuan Province. The accuracy of the models was evaluated using RMSE, MAE, MAPE and r (correlation coefficient) as the parameters. The spatio-temporal evolution characteristics of Sichuan air quality information entropy were evaluated through Mann-Kendall nonparametric test, and calculation of individual influence index, closeness centrality index, betweenness centrality index and eigenvector centrality index. The results can be summarized as follows: (1) The weights of the factors that affect air quality were: PM2.5, PM10, SO2, CO, NO2, O3, monthly average precipitation, and relative humidity, in a descending order; (2) The EWM-LSTM model had the highest accuracy in predicting air quality, with RMSE, MAE, MAPE and r values of 0.012, 0.011, 1.400 and 0.942, respectively. (3) The air quality of the 21 cities in Sichuan Province exhibited significant seasonal variation with high air quality observed in winter and low air quality observed in summer. The Mann-Kendall non-parametric test results showed a significant increase in air quality in 2017. (4) The air quality information entropy in Sichuan Province increased from southwest to northeast with Liangshan Prefecture, Meishan, Ziyang, Zigong, Luzhou, and Guangyuan cities (individual impact index was above 0.490) having the highest impact on air quality. The present findings provide a basis for air quality prediction and provide information for development of strategies to minimize and manage air pollution.
