arXiv 2505.003022025-05-01cs.LG

用于多变量时间序列预测的时序注意力演化图卷积网络

Temporal Attention Evolutional Graph Convolutional Network for Multivariate Time Series Forecasting

College of Information Science and Engineering, China University of Petroleum Beijing（信息科学与工程学院，中国石油大学（北京））；Beijing Key Laboratory of Petroleum Data Mining, China University of Petroleum Beijing（北京石油数据挖掘重点实验室，中国石油大学（北京））

Xinlong Zhao, Liying Zhang, Tianbo Zou, Yan Zhang

备注：: 13 pages, 7 figures

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出TAEGCN模型，通过结合因果时序卷积、多头自注意力机制和动态图结构构建，有效捕捉多变量时间序列的时空特征，在METR-LA和PEMS-BAY数据集上取得优越预测性能。

摘要：

多变量时间序列预测通过利用历史数据来预测未来状态，从而辅助决策过程。多变量时间序列中的每个数据节点包含多个维度的序列。这些节点表现出相互依赖的关系，形成图结构。虽然现有的预测方法通常假设固定的图结构，但许多现实场景涉及动态图结构。此外，在不同时间尺度上观察到的时间序列之间的相互作用差异显著。为了通过捕捉精确的时空特征来提高预测精度，本文提出了时序注意力演化图卷积网络（TAEGCN）。这种新方法不仅集成了因果时序卷积和多头自注意力机制来学习节点的时间特征，还基于这些时间特征构建动态图结构，以保持空间特征变化与时间序列的一致性。TAEGCN巧妙地捕捉了数据内的时间因果关系和隐藏的空间依赖关系。此外，TAEGCN结合了一个统一神经网络，无缝集成这些组件以生成最终预测。在两个公共交通网络数据集METR-LA和PEMS-BAY上进行的实验结果表明，所提出的模型具有优越的性能。

英文摘要：

Multivariate time series forecasting enables the prediction of future states by leveraging historical data, thereby facilitating decision-making processes. Each data node in a multivariate time series encompasses a sequence of multiple dimensions. These nodes exhibit interdependent relationships, forming a graph structure. While existing prediction methods often assume a fixed graph structure, many real-world scenarios involve dynamic graph structures. Moreover, interactions among time series observed at different time scales vary significantly. To enhance prediction accuracy by capturing precise temporal and spatial features, this paper introduces the Temporal Attention Evolutional Graph Convolutional Network (TAEGCN). This novel method not only integrates causal temporal convolution and a multi-head self-attention mechanism to learn temporal features of nodes, but also construct the dynamic graph structure based on these temporal features to keep the consistency of the changing in spatial feature with temporal series. TAEGCN adeptly captures temporal causal relationships and hidden spatial dependencies within the data. Furthermore, TAEGCN incorporates a unified neural network that seamlessly integrates these components to generate final predictions. Experimental results conducted on two public transportation network datasets, METR-LA and PEMS-BAY, demonstrate the superior performance of the proposed model.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。