.png){kind=icon}
Return
MGCRN: Missing-aware Graph Convolutional Recurrent Network for spatio-temporal forecasting
.png){kind=icon}
PRE
AI
DOI：10.1007/s10586-026-06604-w.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Spatio-temporal forecasting (STF) plays a crucial role in real-world applications. However, most existing methods assume complete data and struggle to handle missing values. Additionally, they are less effective for utilizing spatio-temporal heterogeneity. To address these challenges, we propose a Missing-aware Graph Convolutional Recurrent Network (MGCRN) based on an attention mechanism, which accounts for missing data and spatial-temporal heterogeneity. The model consists of two key components: spatio-temporal interactive perception attention and adaptive graph convolutional network. Specifically, we design the interactive perception attention to capture the spatio-temporal dependencies, effectively perceiving missing data and modeling contextual information. Additionally, our proposed adaptive graph convolutional network utilizes the inherent heterogeneity of spatio-temporal data to construct dynamic graphs and reconstruct a reasonable spatio-temporal dependency. Experimental results demonstrate that our model maintains high forecasting accuracy under high missing-rate scenarios and significantly outperforms existing spatio-temporal forecasting methods on four real-world datasets.
Keywords:
Spatio-temporal forecasting
Missing data
Interactive perception attention
Adaptive graph convolution
Journal
IF:
2.9
Papers:
231
Citations:
1.1K
Organization
No organization information available
Cited Papers
ShareSaveGinAR+: A Robust End-to-End Framework for Multivariate Time Series Forecasting With Missing Values
PREAI
ShareSaveA new hybrid ensemble deep reinforcement learning model for wind speed short term forecasting.png)
ENERGY
ENERGYIF9.4
PREAI
ShareSaveSpatio-Temporal Forecasting: A Survey of Data-Driven Models Using Exogenous DataIEEE ACCESS
OAAI
IEEE ACCESSIF3.6
OAAIShareSaveShareSave
View more