Air-Quality Prediction Based on the ARIMA-CNN- LSTM Combination Model optimized by Dung Beetle Optimizer

The air pollution problem seriously affects economic development and people's health, and an efficient and accurate forecasting model of air quality will help manage air pollution problems. The comparison models chosen by other scholars are often based on derivative models of the proposed model and do not comprehensively compare other types of models with limited accuracy. In this paper, we establishes a combined ARIMA-CNN-LSTM model to predict the air quality index accurately. The model mainly consists of two parts: Using the ARIMA model to fit the linear part of the data and using the CNN-LSTM model to fit the nonlinear part of the data to avoid the problem of blindness in the CNN-LSTM hyperparameter setting. To avoid the dilemma of blindness in a CNN-LSTM hyperparameter setting, this article uses the Dung Beetle Optimizer tool to find the hyperparameters of a CNN-LSTM model, determine the best hyperparameters, and check the accuracy of the model. The proposed model is compared with other widely used models. The results show that the Dung Beetle Optimizer can effectively search for the optimal hyperparameters of the model and can solve the problem of blindness in setting the hyperparameters of the model. And the optimized ARIMA-DBO-CNN-LSTM model has higher predictive accuracy with stronger adaptability in predicting the three cities.