Research On Air Quality Index Forecast Based On SSA-GRU

Along with water and land,air is the basic necessity of life.However,due to the increasing urbanization and industrialization,the air quality around the world is deteriorating.The problem of air pollution has become a prominent problem for the public and academia.It is very important to establish a high-precision and high-performance air quality prediction model.Due to the randomness and non-stationarity of the AQI raw data,it is difficult to accurately predict the air quality index,therefore high-precision prediction is currently a challenging problem.In order to improve the accuracy of AQI prediction,this thesis proposes a new SSA-GRU method that the Gated Recurrent Unit(GRU)network model is optimized by the Salp Swarm Algorithm(SSA)to establish a prediction model.First of all,compare the performance of four benchmark models(ARIMA,BPNN,SVR,GRU)in short-term forecasting,and select the GRU model with the best forecasting performance.It can be concluded that the GRU model can better capture the time dependency in a short time.Then,the SSA algorithm optimizes the number of neurons in the two layers of the GRU model and improves the generalization ability of the model.Finally,seven methods(ARIMA,SVR,BPNN,GRU,GA-GRU,PSO-GRU and SSA-GRU)are used to predict the AQI level of Chongqing and Chengdu in China.And compare and analyze the prediction results of SSA-GRU and other models.Experimental results show that the model has good characteristics in MAE,RMSE,MAPE and~2,which has the highest accuracy and the best prediction performance also.At the same time,this model predicts AQI more accurately than other models in this article and the predicted value is closer to the true value,so the prediction effect is better.