Air Quality Index Forecasting Via Genetic Algorithm-Based Improved Extreme Learning Machine

Air quality has always been one of the most important environmental concerns for general public and society. Using machine learning algorithms Quality Index (AQI) prediction is helpful analysis future air trends from a macro perspective. When conventionally using single model to predict quality, it challenging achieve good outcome under various AQI fluctuation trends. In order effectively address this problem, genetic algorithm-based improved extreme (GA-KELM) method enhanced. First, kernel introduced produce matrix which replaces output hidden layer. To issue conventional limit where number nodes random generation thresholds weights lead degradation network ability, algorithm then used optimize layers machine. The thresholds, weights, root mean square error are define fitness function. Finally, least squares applied compute model. verify predictive ability GA-KELM, based on collected basic data long-term forecast at monitoring point in city China, optimized ( SO 2 , xmlns:xlink="http://www.w3.org/1999/xlink">NO xmlns:xlink="http://www.w3.org/1999/xlink">PM xmlns:xlink="http://www.w3.org/1999/xlink">10 xmlns:xlink="http://www.w3.org/1999/xlink">CO xmlns:xlink="http://www.w3.org/1999/xlink">O xmlns:xlink="http://www.w3.org/1999/xlink">3 xmlns:xlink="http://www.w3.org/1999/xlink">PM xmlns:xlink="http://www.w3.org/1999/xlink">2.5 concentration AQI), with comparative experiments convolutional neural networks, recurrent networks long short-term memory. results show that proposed trains faster makes more accurate predictions.