Multi-Dataset Training for Improved Accuracy in Spatio-Temporal Problems: An Explainable Analysis
Mathematics, vol. 14, no. 5, article 908, 2026
← Research programme
02
Research area
Graph learning for sensor systems
Graph neural networks that capture relationships across distributed environmental sensors, cities and global observational grids.
Rationale
The research problem.
Environmental observations are connected by geography, transport processes and shared context. Graph learning makes those relationships explicit, from local monitoring stations to global grids, and combines them with temporal models that represent how signals evolve.
Research questions
Questions that guide this area.
- 01
Which graph structure captures both local and long-range environmental dependencies?
- 02
How can node representations transfer across related sensor networks?
- 03
Which spatial and temporal components account for an observed performance gain?
Methods, systems and sources
Concrete research elements.
- GraphSAGE
- Spectral learning
- Message passing
- Graph transformers
- Graph neural networks
- Sensor networks
- Forecasting
Evidence
Publications in this area.
Global forecasting of aerosol optical depth through a deep learning spatiotemporal modeling
International Journal of Environmental Science and Technology, vol. 23, no. 1, article 69, 2026
Spatio-temporal graph neural network for inter-city air quality forecasting
International Journal of Environmental Science and Technology, vol. 23, no. 1, article 63, 2026
Noise Pollution Prediction in a Densely Populated City Using a Spatio-Temporal Deep Learning Approach
Applied Sciences, vol. 15, no. 10, article 5576, 2025
Global forecasting of carbon concentration through a deep learning spatiotemporal modeling
Journal of Environmental Management, vol. 371, article 122922, 2024