Nature-Inspired Meta-Heuristic Optimization Algorithms Integrated With Ensemble Machine Learning Techniques for Particulate Matter 2.5 Modeling

Particulate Matter 2.5 is a major air pollutant that can deeply penetrate the respiratory system and enter the bloodstream when inhaled. The current study employs the use of both Nature inspired Meta-heuristic optimization algorithms and Ensemble Machine learning techniques for the estimation of Particulate Matter 2.5 value using Sulfur dioxide, Nitrogen Dioxide, Respiratory suspended particulate matter. Prior to dwelling into the modeling step, various pre-analysis techniques were conducted for data clean up and to understand the behavior of the data. The quantitative performance results obtained from the Metaheuristic algorithms indicate that Artificial neural network-Particle swarm optimization outperformed all the other techniques including Support vector regression-Bayesian optimization, Emotional neural network-Genetic algorithms and linear regression. Furthermore, the quantitative outcomes indicate that Artificial neural network-Particle swarm optimization has the ability to improve the performance of the other techniques up to 80.4% and 73.2% in the calibration and validation phases respectively. Moreover, recent visualizations such as Fan plot and Bump chart were used in ranking the performance results obtained in Particulate Matter 2.5 estimation. Moreover, the Neural network ensemble technique equally showed superior potentials over the Simple Average ensemble technique. To conclude, the quantitative and visualized performances of both the Metaheuristic algorithms and the ensemble paradigms indicate their importance in modeling of Particulate Matter 2.5 pollution, which requires concerted efforts at the local and international levels to mitigate its effects and improve air quality on a global scale.