STAMImputer: Spatio-Temporal Attention MoE for Traffic Data Imputation
Authors: Yiming Wang, Hao Peng, Senzhang Wang, Haohua Du, Chunyang Liu, Jia Wu, Guanlin Wu
IJCAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments are conducted on four traffic datasets for evaluation. The result shows STAMImputer achieves significantly performance improvement compared with existing SOTA approaches. Our codes are available at https://github. com/Ring BDStack/STAMImupter. Extensive comparative experiments on four real-world benchmark datasets demonstrate that our STAMImputer model outperforms others in traffic data imputation. The imputation results on the four benchmarks are shown in Table 1. To evaluate the robustness of the model on highly sparse data, we add imputation experiments with varying degrees of sparsity. The results are shown in Figure 2. We also perform ablation studies to verify the significance of the STAMImputer framework designs. The imputation results of the ablation experiments are shown in Figure 3. |
| Researcher Affiliation | Collaboration | Yiming Wang1, Hao Peng*1, 2, 3, Senzhang Wang*4, Haohua Du*1, Chunyang Liu5, Jia Wu6, and Guanlin Wu*7 1School of Cyber Science and Technology, Beihang University 2Hangzhou Innovation Institute of BUAA, Hangzhou, China 3Department of Computer Science and Technology, Shantou University 4School of Computer Science and Engineering, Central South University 5Didi Chuxing 6Department of Computing, Macquarie University 7National University of Defense Technology, Changsha, China EMAIL, EMAIL, EMAIL, EMAIL, EMAIL. |
| Pseudocode | No | The paper describes the methodology using textual explanations and mathematical formulas, but it does not include a distinct block explicitly labeled as 'Pseudocode' or 'Algorithm'. |
| Open Source Code | Yes | Our codes are available at https://github. com/Ring BDStack/STAMImupter. |
| Open Datasets | Yes | Extensive experiments are conducted on four traffic datasets for evaluation. The detailed spatio-temporal feature information of the selected benchmark datasets is shown in Table 2. Table 2: Benchmark datasets details. Pems D8, SZ-Taxi, Di Di-SZ, NYC-Taxi. |
| Dataset Splits | No | The paper specifies how missing data is generated for point missing and block missing patterns (e.g., "random missing rates of 25% and 60%", "failure probabilities of 0.2% and 1%"), but it does not explicitly state the training, validation, and test splits (e.g., 80/10/10%) for the datasets themselves. It focuses on the missing data scenarios for evaluation. |
| Hardware Specification | Yes | All experiments are conducted on a server with an Intel(R) Xeon(R) Platinum 8336C CPU operating at 2.30GHz and an NVIDIA A800 GPU with 80GB of memory for the above models and datasets. |
| Software Dependencies | No | The paper does not provide specific version numbers for any software dependencies or libraries used in the experiments. It implicitly refers to general deep learning frameworks but without concrete version details. |
| Experiment Setup | No | The paper describes the model architecture and general experimental settings like missing patterns and evaluation metrics, but it does not specify concrete hyperparameter values (e.g., learning rate, batch size, number of epochs, optimizer details) for training the models. |