Graph correlated attention recurrent neural network for multivariate time series forecasting

Multivariate time series(MTS) forecasting is an urgent problem for numerous valuable applications. At present, attention-based methods can relieve recurrent neural networks' limitations in MTS forecasting that are hard to focus on key information and capture long-term dependencies, but they fail to learn the time-varying pattern based on the reli-able interaction. To reinforce the memory ability of key features across time, we propose a Graph Correlated Attention Recurrent Neural Network(GCAR). GCAR first nests Feature -level attention in the graph attention module to complement external feature representa-tions on the extraction of multi-head temporal correlations. Then Multi-level attention is designed to add target factors' impact on the selection of external correlation and achieve a fine-grained distinction of external features' contribution. To better capture different ser-ies' continuous dynamic changes, two parallel LSTMs are respectively applied to learn his-torical target series and external feature representations' temporal dependencies. Finally, a fusion gate is employed to balance their information conflicts. The performance of GCAR model is tested on 4 datasets, and results show GCAR model performs the most stable and greatest predictive accuracy as the increasing of predicted horizons compared with state-of-the-art models even if the multivariate time series present strong volatility and randomness.(c) 2022 Published by Elsevier Inc.

Automated summary

Multivariate time series (MTS) forecasting is a pressing problem, and attention-based methods can alleviate the limitations of recurrent neural networks. However, they fail to learn time-varying patterns. To address this, a GCAR model is proposed, which exhibits stability and achieves the highest predictive accuracy in experiments.