Multibranch adaptive fusion graph convolutional network for traffic flow prediction
Urban road networks have complex spatial and temporal correlations, driving a surge of research interest in spatial-temporal traffic flow prediction. However, prior approaches often overlook the temporal-scale differentiation of spatial-temporal features, limiting their ability to extract complex st...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2023
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/171178 |
| _version_ | 1826109622984900608 |
|---|---|
| author | Zan, Xin Lam, Jasmine Siu Lee |
| author2 | School of Civil and Environmental Engineering |
| author_facet | School of Civil and Environmental Engineering Zan, Xin Lam, Jasmine Siu Lee |
| author_sort | Zan, Xin |
| collection | NTU |
| description | Urban road networks have complex spatial and temporal correlations, driving a surge of research interest in spatial-temporal traffic flow prediction. However, prior approaches often overlook the temporal-scale differentiation of spatial-temporal features, limiting their ability to extract complex structural information. In this work, we design the multibranch adaptive fusion graph convolutional network (MBAF-GCN) that explicitly exploits the prior spatial-temporal characteristics at different temporal scales, and each branch is responsible for extracting spatial-temporal features at a specific scale. Besides, we design the spatial-temporal feature fusion (STFF) module to refine the prediction results. Based on the multibranch complementary features, the module adopts a coarse-to-fine fusion strategy, incorporating different spatial-temporal scale features to obtain recalibrated prediction results. Finally, we evaluate the MBAF-GCN using two real-world traffic datasets. Experimentally, the newly designed multibranch can efficaciously utilize the prior information of different temporal scales. Our MBAF-GCN achieved better performance in the comparative model, indicating its potential and validity. |
| first_indexed | 2024-10-01T02:21:15Z |
| format | Journal Article |
| id | ntu-10356/171178 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T02:21:15Z |
| publishDate | 2023 |
| record_format | dspace |
| spelling | ntu-10356/1711782023-10-20T15:33:12Z Multibranch adaptive fusion graph convolutional network for traffic flow prediction Zan, Xin Lam, Jasmine Siu Lee School of Civil and Environmental Engineering Engineering::Civil engineering Adaptive Fusion Convolutional Networks Urban road networks have complex spatial and temporal correlations, driving a surge of research interest in spatial-temporal traffic flow prediction. However, prior approaches often overlook the temporal-scale differentiation of spatial-temporal features, limiting their ability to extract complex structural information. In this work, we design the multibranch adaptive fusion graph convolutional network (MBAF-GCN) that explicitly exploits the prior spatial-temporal characteristics at different temporal scales, and each branch is responsible for extracting spatial-temporal features at a specific scale. Besides, we design the spatial-temporal feature fusion (STFF) module to refine the prediction results. Based on the multibranch complementary features, the module adopts a coarse-to-fine fusion strategy, incorporating different spatial-temporal scale features to obtain recalibrated prediction results. Finally, we evaluate the MBAF-GCN using two real-world traffic datasets. Experimentally, the newly designed multibranch can efficaciously utilize the prior information of different temporal scales. Our MBAF-GCN achieved better performance in the comparative model, indicating its potential and validity. Nanyang Technological University Published version The authors acknowledge the funding support from project 04SBS000097C120 at Nanyang Technological University, Singapore. 2023-10-16T07:46:42Z 2023-10-16T07:46:42Z 2023 Journal Article Zan, X. & Lam, J. S. L. (2023). Multibranch adaptive fusion graph convolutional network for traffic flow prediction. Journal of Advanced Transportation, 2023, 1-13. https://dx.doi.org/10.1155/2023/8256907 0197-6729 https://hdl.handle.net/10356/171178 10.1155/2023/8256907 2-s2.0-85163545434 2023 1 13 en 04SBS000097C120 Journal of Advanced Transportation © 2023 Xin Zan and Jasmine Siu Lee Lam. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. application/pdf |
| spellingShingle | Engineering::Civil engineering Adaptive Fusion Convolutional Networks Zan, Xin Lam, Jasmine Siu Lee Multibranch adaptive fusion graph convolutional network for traffic flow prediction |
| title | Multibranch adaptive fusion graph convolutional network for traffic flow prediction |
| title_full | Multibranch adaptive fusion graph convolutional network for traffic flow prediction |
| title_fullStr | Multibranch adaptive fusion graph convolutional network for traffic flow prediction |
| title_full_unstemmed | Multibranch adaptive fusion graph convolutional network for traffic flow prediction |
| title_short | Multibranch adaptive fusion graph convolutional network for traffic flow prediction |
| title_sort | multibranch adaptive fusion graph convolutional network for traffic flow prediction |
| topic | Engineering::Civil engineering Adaptive Fusion Convolutional Networks |
| url | https://hdl.handle.net/10356/171178 |
| work_keys_str_mv | AT zanxin multibranchadaptivefusiongraphconvolutionalnetworkfortrafficflowprediction AT lamjasminesiulee multibranchadaptivefusiongraphconvolutionalnetworkfortrafficflowprediction |