Lin, Ziqian, et al. "Deepstn+: Context-aware spatial-temporal neural network for crowd flow prediction in metropolis." Proceedings of the AAAI conference on artificial intelligence. Vol. 33. No. 01. 2019.
https://ojs.aaai.org/index.php/AAAI/article/view/3892
DeepSTN+: Context-Aware Spatial-Temporal Neural Network for Crowd Flow Prediction in Metropolis | Proceedings of the AAAI
ojs.aaai.org
Region-level traffic prediction에서 눈여겨볼 Paper.
1. 주요 Contribution은 long-range spatial dependence를 capture한다.
2. POI관련 데이터를 활용하여 locational function을 반영
3. stable한 neural network를 제안하였다.
아키텍쳐는 다음과 같다.
위의 빨강, 초록, 파랑은 Closeness, Period, Trend 정보가 되겠다.
그리고 POI distribution을 [0,1]로 normalize하여 feature로 활용하다.
다양한 conv layer가 활용된다.
다음의 데이터셋을 활용하였다.
https://github.com/FIBLAB/DeepSTN
GitHub - FIBLAB/DeepSTN: Codes for AAAI 2019 DeepSTN+: Context-aware Spatial-Temporal Neural Network for Crowd Flow Prediction i
Codes for AAAI 2019 DeepSTN+: Context-aware Spatial-Temporal Neural Network for Crowd Flow Prediction in Metropolis - GitHub - FIBLAB/DeepSTN: Codes for AAAI 2019 DeepSTN+: Context-aware Spatial-Te...
github.com
'Paper Review' 카테고리의 다른 글
| Hierarchical Graph Convolution Network for Traffic Forecasting (AAAI-21) (0) | 2022.01.17 |
|---|---|
| Fine-grained Urban Flow Prediction (WWW'21) (0) | 2022.01.11 |
| Knowledge Adaption for Demand Prediction based on Multi-task Memory Neural Network (CIKM-20) (0) | 2021.12.16 |
| Discrete Graph Structure Learning for Forecasting Multiple Time Series (0) | 2021.12.14 |
| Traffic Prediction 관련 논문/데이터셋 정리 (0) | 2021.01.15 |
댓글