Spatio-temporal graph neural networks for missing data completion in traffic prediction

Abstract

Missing traffic data completion is a key part of the construction of a smart city. However, due to cost constraints and other reasons, many locations do not have sensors to record traffic data. Most research methods do not systematically consider filling in missing traffic data. This study explores a new spatio-temporal feature extraction layer that includes spatio-temporal feature fusion, graph learning on an adaptive adjacency matrix, and a gated recurrent unit with a mask for missing traffic data completion. This idea is based on a hypothesis: missing data can be inferred from the spatio-temporal features of other nearby recorded sensor nodes. Therefore, we propose an end-to-end traffic model dealing with missing data - missing traffic data completion graph neural networks (MTC-GNN). Experiments demonstrate that the proposed model can learn spatio-temporal patterns and fill in speed from traffic data with various missing ratios and outperform existing models.

Keywords:

• Missing traffic data
• data completion
• graph neural networks

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data and codes availability statement

The code and data of this paper can also be found in our GitHub repository at https://github.com/ucasVachel/MTC-gnn.

Additional information

Notes on contributors

Jiahui Chen

Jiahui Chen is a doctoral student at the Aerospace Information Research Institute (AIR) under the Chinese Academy of Sciences (CAS) in China. His research interests lie in spatio-temporal intelligence and decision support. He contributed to conceptualization, software development, writing, reviewing, editing, and visualization.

Lina Yang

Lina Yang is an Associate Professor in Geoinformatics and leads the Spatial-temporal Knowledge Graph Lab at the Aerospace Information Research Institute (AIR) under the Chinese Academy of Sciences (CAS), China. Her research interests lie in the integration and computation of spatiotemporal big data. She contributed to writing (review and editing), supervision, and project administration.

Yi Yang

Yi Yang is an Associate Professor in Computer Science at the Institute of Automation of the Chinese Academy of Sciences. His research explores the knowledge graph reasoning for spatial information modeling. He contributed to conceptualisation, methodology, review, and supervision.

Ling Peng

Ling Peng is a Professor of Geoinformation Engineering at the Aerospace Information Research Institute (AIR) under the Chinese Academy of Sciences (CAS). Her research interests lie in spatio-temporal data applications. She contributed to supervision and project administration.

Xingtong Ge

Xingtong Ge is a doctoral student at the Aerospace Information Research Institute (AIR) under the Chinese Academy of Sciences (CAS) in China. His research interests lie in cross-domain spatio-temporal knowledge graph construction and spatio-temporal big data situation analysis. She contributed to editing and visualization.