其中,Gconv(⋅)是一个图卷积层。图卷积递归网络(Graph Convolutional Recurrent Network, GCRN)[71] 将LSTM网络与ChebNet [21] 结合在一起。扩散卷积递归神经网络(Diffusion Convolutional Recurrent Neural Network, DCRNN)[72] 将提出的扩散图卷积层(方程18)结合到GRU网络中。此外,DCRNN采用了编码器-解码器框架来...
BigST: Linear Complexity Spatio-Temporal Graph Neural Network for Traffic Forecasting on Large-Scale Road Networks 链接:vldb.org/pvldb/vol17/p1 代码:github.com/usail-hkust/ 作者:Jindong Han, Weijia Zhang, Hao Liu, Tao Tao, Naiqiang Tan, Hui Xiong 关键词:可扩展的交通预测 BigST Sparcle: Bo...
Graph neural networkSpatial-temporal graph neural networkTraffic forecasting plays an important role of modern Intelligent Transportation Systems (ITS). With the recent rapid advancement in deep learning, graph neural networks (GNNs) have become an emerging research issue for improving the traffic ...
This paper introduces a novel application of spatial-temporal graph neural networks (ST-GNNs) to predict groundwater levels. Groundwater level prediction is inherently complex, influenced by various hydrological, meteorological, and anthropogenic factors. Traditional prediction models often struggle with the...
Spatial-Temporal Fusion Graph Neural Module 轻量级深度学习模型通过简单的时空就能提取出隐藏的时空相关性,比如乘法运算。通过与AST F G的多次矩阵乘法,网络中每个节点可以聚合ASG的空间相关性、AT G的时间模式相关性和AT C的自身近相关长时间轴。通过叠加L个图乘法块,可以聚集更复杂的非局部空间依赖。
Li M, Zhu Z (2021) Spatial-temporal fusion graph neural networks for traffic flow forecasting. Proc AAAI Conf Artif Intell 35:4189–4196 Google Scholar Chen Y, Segovia-Dominguez I, Gel YR (2021) Z-gcnets: Time zigzags at graph convolutional networks for time series forecasting. In: Meil...
BigST: Linear Complexity Spatio-Temporal Graph Neural Network for Traffic Forecasting on Large-Scale Road Networks 链接:https://www.vldb.org/pvldb/vol17/p1081-han.pdf 代码:https://github.com/usail-hkust/BigST 作者:Jindong Han, Weijia Zhang, Hao Liu, Tao Tao, Naiqiang Tan, Hui Xiong ...
2) The susceptibility of the existing graph neural networks (GNNs) to oversmoothing reflects an inherent limitation of these networks. As the network layers deepen, all node representations tend to converge to a uniform value, which greatly affects the ability of the employed model to capture long...
Connecting the Dots: Multivariate Time Series Forecasting with Graph Neural Networks(KDD 2020)中对这类方法不同的拉普拉斯矩阵学习方式进行了对比。下表列出了不同的拉普拉斯矩阵生成方法的效果对比,这类方法的一半思路是根据每个节点的Embedding,通过不同的计算方式得到拉普拉斯矩阵中每个元素的值,这个值直接用于后续...
Location-based Graph Neural Network 更新边的特征 更新点的特征 更新全局图的特征 Spatial-temporal Attention Net 对于时间切片 对于空间切片 使用注意力机制,X是3维特征矩阵,对于每一个交易记录 使用时间注意力机制,更新时间切片,在此基础上在使用空间注意力机制 ...