GENETIC ALGORITHM-BASED OPTIMIZATION OF EXTREME LEARNING MACHINE FOR ACCURATE AIR QUALITY INDEX FORECASTING

Abstract

Air Quality Index (AQI) forecasting plays a critical role in mitigating the adverse health effects of air pollution and informing public policy decisions. Traditional models often fall short due to the non-linear and complex nature of air quality data. This research presents a novel approach that leverages a Genetic Algorithm (GA)-based Improved Extreme Learning Machine (ELM) for more accurate AQI forecasting. The proposed method addresses the limitations of standard ELM by optimizing the input weights and biases using a Genetic Algorithm, thereby enhancing the model's predictive capability. The integration of GA with ELM allows for the automatic selection of optimal parameters, which leads to a more robust model that can handle the highly variable nature of air quality data. The model's performance is evaluated using real-world AQI datasets, demonstrating superior accuracy compared to traditional machine learning methods. The improved ELM effectively captures the underlying patterns in the data, providing reliable short-term AQI predictions. This study highlights the importance of model optimization in improving forecasting accuracy, particularly for complex environmental data. The GA-based Improved ELM not only offers better performance but also contributes to the broader field of environmental monitoring and public health by providing timely and accurate AQI forecasts. The results of this research suggest that the proposed method could be a valuable tool for policymakers and environmental agencies aiming to reduce the impact of air pollution on public health.