2021 7th International Conference on Computing and Data Engineering 2021
DOI: 10.1145/3456172.3456210
|View full text |Cite
|
Sign up to set email alerts
|

Spatio-Temporal Multi-Scale Convolutional Network for Traffic Forecasting

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
7
0

Year Published

2022
2022
2024
2024

Publication Types

Select...
3
1

Relationship

0
4

Authors

Journals

citations
Cited by 4 publications
(7 citation statements)
references
References 15 publications
0
7
0
Order By: Relevance
“…DCRNN [2]: Spatial features are captured using diffusion convolution, whose diffusion process is based on random wandering on the graph, and temporal features of the traffic flow are captured using recurrent neural networks.…”
Section: Benchmarking Methodsmentioning
confidence: 99%
See 2 more Smart Citations
“…DCRNN [2]: Spatial features are captured using diffusion convolution, whose diffusion process is based on random wandering on the graph, and temporal features of the traffic flow are captured using recurrent neural networks.…”
Section: Benchmarking Methodsmentioning
confidence: 99%
“…Guo et al [25] further added the spatial and temporal attention mechanisms based on MCSTGCN to extract spatio-temporal features separately, improving the prediction performance of the model. Li et al [2] proposed a diffusion convolution recurrent neural network (DCRNN), which uses a bi-directional random wandering diffusion GCN to extract spatial dependencies. Wu et al [26] proposed multivariate time series forecasting networks (MTGNN), which utilize the intrinsic dependencies between multiple road segments for traffic forecasting.…”
Section: Traffic Flow Forecastingmentioning
confidence: 99%
See 1 more Smart Citation
“…In an alternative approach, the diffusion convolutional recurrent network (DCRNN) is proposed to improve the spatio-temporal modeling capabilities of spectral graph convolution. This model generalized the graph convolution approach to operate on both directed and undirected graphs (Li et al, 2018c). The key contribution of this work is the proposed diffusion convolution operation, which relates the traffic flow stream to a diffusion process and can effectively capture stochastic and noisy data trends.…”
Section: Gcnmentioning
confidence: 99%
“…Large volumes of spatio-temporal (ST) data (e.g., GPS samples, location-based video footages, and remote sensing data) are increasingly collected and studied in diverse domains, including human mobility [118,119], intelligent transportation [120,121], urban planning [122,123], epidemiology [124,125], as well as environmental and climate science [126,127]. While recent advances in Machine Learning (ML), especially Deep Learning (DL) see great benefits in leveraging big ST data to facilitate model training and inference, transforming the raw ST data into ML ingestible features, however, still faces major challenges.…”
Section: Introductionmentioning
confidence: 99%