Responsive Dynamic Graph Disentanglement for Metro Flow Forecasting
Authors: Qiang Gao, Zizheng Wang, Li Huang, Goce Trajcevski, Guisong Liu, Xueqin Chen
AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our experiments conducted on three real-world metro passenger flow datasets demonstrate that the proposed Re Dy Net outperforms several representative baselines. ... Datasets. We select three real-world metro flow datasets: Beijing Metro (Zhang et al. 2020), Shanghai Metro (Liu et al. 2020), and Hangzhou Metro (Liu et al. 2020). ... Implementations. Our Re Dy Net is implemented with Py Torch, accelerated by an NVIDIA RTX 4090. ... Metrics. We use three commonly used evaluation protocols, including mean absolute error (MAE), root mean squared error (RMSE), and mean absolute percentage error (MAPE). ... Overall Performance. Table 1 reports the results of all methods on three datasets... Ablation Study. We now investigate the impact of each module design in Re Dy Net. Correspondingly, we yield five variants of Re Dy Net... |
| Researcher Affiliation | Collaboration | 1School of Computing and Artificial Intelligence, Southwestern University of Finance and Economics, Chengdu, China 2Engineering Research Center of Intelligent Finance, Ministry of Education, Chengdu, China 3Iowa State University, Iowa, USA 4Kash Institute of Electronics and Information Industry, Kashgar, China EMAIL, EMAIL, EMAIL, EMAIL, EMAIL, EMAIL |
| Pseudocode | No | The paper describes the methodology using mathematical equations and detailed textual explanations within sections like "2. Problem and Methodology", "2.1 Contextual Spatial Embedding (CSE)", "2.2 Redundancy Context Disentanglement (RCD)", "2.3 Responsive Dynamic Learning (RDL)", and "2.4 Task Learning Task Adaption". However, it does not include any explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | Yes | For reproducibility, the source codes are available at https://github.com/wangzz-yyzz/Re Dy Net. |
| Open Datasets | Yes | Datasets. We select three real-world metro flow datasets: Beijing Metro (Zhang et al. 2020), Shanghai Metro (Liu et al. 2020), and Hangzhou Metro (Liu et al. 2020). |
| Dataset Splits | Yes | We follow the standard dataset split manner by dividing the original into training, validation, and testing sets with a ratio of 7:1:2. |
| Hardware Specification | Yes | Our Re Dy Net is implemented with Py Torch, accelerated by an NVIDIA RTX 4090. |
| Software Dependencies | No | Our Re Dy Net is implemented with Py Torch, accelerated by an NVIDIA RTX 4090. While PyTorch is mentioned, a specific version number is not provided, nor are other key software dependencies with their versions. |
| Experiment Setup | Yes | Implementations. Our Re Dy Net is implemented with Py Torch, accelerated by an NVIDIA RTX 4090. de is set to 128, dr is set to 32 and cw is 6. go in RDL is set to 256, and the order of the Chebyshev polynomials k is 4. We choose Adam as our optimizer with up to 300 epochs. b is set to 16, α is 0.01 and the initial learning rate is 0.003. All parameters were determined through grid search to ensure optimal performance. |