DeepPM_2.5: PM_2.5 Prediction Under Dynamic and Heterogeneous Conditions With Contrastive Learning and Spatiotemporal Graph Convolution (opens in new tab)

Air quality prediction is closely tied to daily life and has attracted significant attention from both governmental agencies and the general public. However, the intricate spatiotemporal heterogeneity embedded in real-world atmospheric circulation, coupled with the impact of diverse dynamic factors, renders accurate air quality prediction highly challenging. To address this issue, we present a novel deep learning framework for PM<sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink=...