Air Quality Index Prediction Using An Enhanced Extreme Learning Machine Based On Genetic Algorithms

Abstract

Security of human wellbeing and heading of ecological arrangement rely basically upon air quality prediction. Precise expectation of air quality changes is quite difficult for the vast majority traditional single-model frameworks. This work answers with areas of strength for a framework utilizing state of the art machine learning draws near. We examine many models like "support Vector Relapse (SVR), Genetic Algorithm-Enhanced Extreme Learning Machine (GA-KELM), and Deep Belief Network with Back-Propagation (DBN-BP)" from a near perspective. To further develop expectation precision considerably more, we likewise recommend including a deep learning design called “bidirectional long short-term memory (BiLSTM)”. Through broad trial and error and assessment, we show that BiLSTM displays lower “Root Mean Square Error (RMSE) and Mean Squared Error (MSE)” values, so astounding current models. In addition, we expand the presentation of BiLSTM by adding GA-KELM, accordingly further developing its prescient powers considerably more. Aside from giving better accuracy in air quality expectation, the recommended half and half model assists with directing general wellbeing efforts and contamination control approaches through informed choices. This study emphasizes the need to explore approaches to conceptually address important biological issues in air quality monitoring and the potential improvement of machine learning outcomes.