To predict air quality (PM2.5 concentrations, et al), many parametric regression models have been developed, while deep learning algorithms are used less often. And few of them takes the air pollution emission or spatial information into consideration or predict them in hour scale. In this paper, we proposed a spatialtemporal GRU-based prediction framework incorporating ground pollution monitoring (GPM), factory emissions (FE), surface meteorology monitoring (SMM) variables to predict hourly PM2.5 concentrations. The dataset for empirical experiments was built based on air quality monitoring in Shenyang, China. Experimental results indicate that our method enables more accurate predictions than all baseline models and by applying the convolutional processing to the GPM and FE variables notable improvement can be achieved in prediction accuracy.

Decrease in air quality is one of the most crucial threats to human health. There is an imperative and necessary need for more accurate air quality prediction. To meet this need, we propose a novel long short-term memory-based deep random subspace learning (LSTM-DRSL) framework for air quality forecasting. Specifically, we incorporate real-time pollutant emission data into the model input. We also design a spatial-temporal analysis approach to make good use of these data. The prediction model is developed by combining random subspace learning with a deep learning algorithm in order to improve the prediction accuracy. Empirical analyses based on multiple datasets over China from January 2015 to September 2017 are performed to demonstrate the efficacy of the proposed framework for hourly pollutant concentration prediction at an urban-agglomeration scale. The empirical results indicate that our framework is a viable method for air quality prediction. With consideration of the regional scale, the LSTM-DRSL framework performs better at a relatively large regional scale (around 200–300 km). In addition, the quality of predictions is higher in industrial areas. From a temporal point of view, the LSTM-DRSL framework is more suitable for hourly predictions.