An advanced deep learning predictive model for air quality index forecasting with remote satellite-derived...

Area of Science:

• Environmental Science
• Computer Science
• Data Science

Background:

• Air pollution poses a significant environmental and public health threat, particularly in developing nations.
• Accurate forecasting of the Air Quality Index (AQI) is crucial for effective environmental management and health risk mitigation.

Purpose of the Study:

• To introduce and evaluate a novel hybrid deep learning model for precise AQI forecasting in a sub-tropical region.
• To assess the model's performance against benchmark methods using real-world data.

Main Methods:

• Development of a hybrid deep learning model (CLSTM-BiGRU) integrating Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and Bi-directional Gated Recurrent Unit (BiGRU).
• Utilization of nineteen remotely sensed predictor variables.
• Optimization of the model using the Grey Wolf Optimization (GWO) algorithm.

Main Results:

• The CLSTM-BiGRU model demonstrated superior performance in AQI forecasting compared to benchmark models.
• The model achieved lower forecasting errors and a high correlation coefficient (approximately 1).

Conclusions:

• The hybrid CLSTM-BiGRU model offers a feasible and potent solution for precise AQI forecasting.
• This model has significant potential for air quality monitoring and public health risk mitigation in urban areas.