Abstract
Missing observations in air quality sensor networks can reduce the reliability of environmental monitoring systems. This paper proposes GCN-KAN, a hybrid model combining graph convolutional networks and Kolmogorov–Arnold Networks for PM data imputation by exploiting spatial and nonlinear temporal dependencies. Experiments conducted on multiple missing-data scenarios demonstrate improved reconstruction accuracy and robustness compared with baseline methods.
Paper Type
Poster
DOI
10.62036/ISD.2026.100
Spatio-temporal imputation of missing air quality data using a hybrid GCN–KAN model
Missing observations in air quality sensor networks can reduce the reliability of environmental monitoring systems. This paper proposes GCN-KAN, a hybrid model combining graph convolutional networks and Kolmogorov–Arnold Networks for PM data imputation by exploiting spatial and nonlinear temporal dependencies. Experiments conducted on multiple missing-data scenarios demonstrate improved reconstruction accuracy and robustness compared with baseline methods.
Recommended Citation
Bernacki, J., Scherer, R., Nowicki, R., Napoli, C., Wei, W. & Hazra, S.(2026). Spatio-temporal imputation of missing air quality data using a hybrid GCN–KAN model. In M. Valenta, B. Mannová, R. Pergl, A. Przybylek, M. Lang, H. Linger, C. Schneider, N. Iivari, & E. Insfran (Eds.), Making ISD Sustainable: Reloaded with AI and Automation (ISD2026 Proceedings). Prague, Czech Republic: Czech Technical University in Prague. ISBN: 978-80-01-07585-2. https://doi.org/10.62036/ISD.2026.100