Air quality index prediction with optimisation enabled deep learning model in IoT application.

Area of Science:

• Environmental Science
• Computer Science
• Data Science

Background:

• Industrial and urban development have led to significant air pollution, impacting human health and the atmosphere.
• Accurate measurement of the Air Quality Index (AQI) is crucial and depends on complex environmental factors like emissions and chemical reactions.

Purpose of the Study:

• To develop an Internet of Things (IoT)-based Deep Learning (DL) technique for predicting air quality.
• To enhance the efficiency of data transmission and the accuracy of air quality predictions.

Main Methods:

• Implemented an IoT simulation with a novel Tangent Two-Stage Algorithm (TTSA) for efficient data routing to the Base Station (BS).
• Utilized Z-score normalization and feature indicator extraction for data preprocessing.
• Employed a Deep Feedforward Neural Network (DFNN) optimized by the Fractional Tangent Two-Stage Optimisation (FTTSA) for AQI prediction.

Main Results:

• The TTSA routing mechanism achieved superior performance with energy consumption of 0.979J, time of 0.025s, and distance of 0.196m.
• The DFNN model demonstrated high accuracy in AQI prediction, with RMSE of 0.602, R-squared of 0.598, MSE of 0.362, and MAPE of 0.456.

Conclusions:

• The proposed IoT-based DL technique effectively predicts air quality.
• The integration of TTSA and FTTSA optimizes the system for efficient data handling and accurate AQI forecasting.