GCGNet uses a variational generator, graph structure aligner, and graph refiner to jointly capture temporal and channel correlations in time series forecasting with exogenous variables, outperforming baselines on 12 real-world datasets.
Tri- former: Triangular, variable-specific attentions for long sequence multivariate time series fore- casting
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
GCGNet: Graph-Consistent Generative Network for Time Series Forecasting with Exogenous Variables
GCGNet uses a variational generator, graph structure aligner, and graph refiner to jointly capture temporal and channel correlations in time series forecasting with exogenous variables, outperforming baselines on 12 real-world datasets.