Air quality prediction model based on mRMR-RF feature selection and ISSA-LSTM

Area of Science:

• Environmental Science
• Data Science
• Machine Learning

Background:

• Severe air pollution presents significant risks to public safety and health.
• Accurate air quality forecasting is essential for effective pollution control and public guidance.
• Existing single-module machine learning models face challenges with long training times and suboptimal prediction accuracy.

Purpose of the Study:

• To develop a novel machine learning approach for enhanced Air Quality Index (AQI) prediction.
• To improve the accuracy and efficiency of air quality forecasting models.

Main Methods:

• A hybrid model integrating Random Forest (RF) with Minimum Redundancy Maximum Relevance (mRMR) for influential variable selection.
• An Improved Sparrow Search Algorithm (ISSA) for optimizing Long Short-Term Memory (LSTM) network hyperparameters.
• Utilizing the optimized LSTM model for predicting AQI concentrations.

Main Results:

• The proposed ISSA-LSTM model demonstrated superior performance compared to other models.
• The model achieved higher prediction accuracy, indicated by lower Root Mean Square Error (RMSE) and higher R-squared (R²) values.
• Validation across different time steps confirmed the model's robust predictive performance with minimal errors.

Conclusions:

• The ISSA-LSTM model offers a viable and effective solution for accurate air quality index forecasting.
• This approach addresses limitations of traditional models, providing a more reliable tool for environmental monitoring and public health protection.