Medium and long-term trend prediction of urban air quality based on deep learning

In order to overcome the low accuracy of traditional air quality change prediction methods, this paper designs a medium and long-term prediction method of urban air quality change trend based on deep learning. The deep learning network is constructed, and the air quality prediction process is designed by using the deep learning algorithm to optimise the air quality prediction model. The deep belief network is initialised by unsupervised training, and the data is supervised by back propagation algorithm. By continuously optimising the network weights to avoid the network falling into local optimum, the medium and long-term accurate prediction of air quality change trend can be realised. The experimental results show the AQI index value of the prediction results of the model has a high fitting degree with the actual value, and the evaluation values of RMSE, MAE, MSE and SMAPE are 2.608%, 2.613%, 2.07% and 0.9513 respectively, which proves the effectiveness of the method.