Air Quality Index Predictions with a Hybrid Forecasting Model: Combining Series Decomposition and Deep Learning Techniques

As China's industrialization process and people's living standards continue to improve, the importance placed on air quality has gradually increased. Precise prediction of the Air Quality Index (AQI) is vital for proactively mitigating the risks posed by air pollution. However, air quality levels can fluctuate significantly over a certain period. Even the interval models used in recent research face the challenge of not being able to retain essential information regarding air quality status. The prediction of the Air Quality Index (AQI) is crucial for controlling air pollution. Furthermore, while pursuing longer prediction times, the accuracy of the prediction must also be ensured. This paper proposes a model based on the fusion of EEMD-CEEMDAN and LSTM. EEMD-CEEMDAN is used to decompose the raw data to reduce its complexity, improve the accuracy and stability of data analysis, and then LSTM is used to predict the AQI. The experiments show that the model proposed in this paper demonstrates excellent experimental results on the real-world collected weather quality dataset.