Research on prediction method of Atmospheric carbon monoxide Concentration based on Deep Learning

In order to solve the problems existing in the existing atmospheric environment prediction models in accurately predicting the change trend of carbon monoxide (CO) concentration in the atmosphere, Sparrow optimization algorithm (SSA) is proposed to optimize the CO concentration prediction model of convolutional neural network (CNN) and bidirectional long Short-Term memory neural network (BILSTM). The SSA algorithm was used to optimize the hyperparameters in the CNN-BILSTM model to confirm the best model. Then, the feature extraction of CO concentration data is carried out by CNN, and the time series modeling is carried out by BILSTM, and finally the CO concentration data prediction results in the next period of time are obtained. The results show that compared with the CNN-BILSTM model, the MAE value of the SSA-CNN-BILSTM model is reduced by 0.0083, the MAPE value is reduced by 2.142%, and the RMSE value is reduced by 0.01. SSA-CNN-BILSTM can predict the change trend of CO concentration more accurately, and can better adapt to complex atmospheric environmental conditions.