Prediction of Air Quality in Major Cities of China by Deep Learning

With global industrialization, air pollution is becoming a critical issue that threatens human health. The World Health Organization (WHO) estimated that air pollution kills several million people worldwide each year. Researchers from various areas and governments and enterprises have invested many resources in investigating and reducing air pollution. Air Quality Index (AQI) is one of the essential indexes indicating air quality or the level of air pollution. A new dataset, including hourly AQI information recorded by 1,615 observation sites covering China from 2015 to 2019, is constructed. Several methods, including linear model and state-of-art techniques, such as Back Propagation Neural Network (BPNN), Convolutional Neural Networks (CNN), Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM), and Bi-directional Long Short-Term Memory (BiLSTM), are adopted to forecast hourly AQI. The performance of these techniques is evaluated, and experiments show that the BiLSTM gives the best performance.