Air Quality Predictive Analysis using Empirical Mode Decomposition with Adaptive Noise-Bee Colony Optimization

One of the most harmful pollutants as PM2.5 (particulate matter with a diameter of less than 2.5 micrometres) may reach deep into the respiratory system, air pollution gravely damages human health and the ecosystem. Accurate and fast PM2.5 concentration monitoring is essential for both environmental protection and good management of public health. Concerning data accuracy, scalability, and reactivity, typical methods of air quality monitoring fail. Recent advancements in big data analytics and artificial intelligence (AI) offer new directions for enhancing air quality monitoring systems. Effective prediction of PM2.5 concentrations in air quality monitoring is challenged by the complex, dynamic character of atmospheric pollutants and ambient variables. The enormous variation in air quality data and the need of adaptive, real-time forecasting test numerous present methods. Much desired for is a more efficient model able to control noisy, non-linear data while increasing forecast accuracy. A new air quality forecasting model based on big data analytics mixed with artificial intelligence—more especially, Empirical Mode Decomposition with Adaptive Noise-Bee Colony Optimisation (EMD-AN-ABC) algorithm is used. The EMD-AN-ABC model separates complex air quality data into intrinsic mode functions (IMFs), therefore reducing noise and spotting significant patterns. The Bee Colony Optimisation approach then enhances the model parameters for increased forecast accuracy. The proposed EMD-AN-ABC model exceeds traditional models on five primary evaluation criteria: Mean absolute error (MAE), root mean squared error (RMSE); coefficient of determination (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>); accuracy, recall; With an MAE of 2.15 μg/m<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup>, an RMSE of 3.08 μg/m<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup>, a R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of 0.92, precision of 89.5%, and recall of 87.3%, the EMD-AN-ABC model, for example, exceeded standard models including ARIMA, SVM, and LSTM by 10–15% across all measures.