search

benedekrozemberczki / pytorch_geometric_temporal

PyTorch Geometric Temporal: Spatiotemporal Signal Processing with Neural Machine Learning Models (CIKM 2021)
MIT License
2.58k stars 367 forks source link

Temporal dataset #244

Open Zhuofeng-Li opened 10 months ago

Zhuofeng-Li commented 10 months ago

Hi, I am sorry to bother you! I have a question about temporal dataset.

In EnglandCovidDatasetLoader , X shape is (num_nodes, time_window feature). Is that right? I think X shape is: (num_nodes, time_window, feature) instead of (num_nodes, time_window feature) (although feature = 1).

Each time_step should has it own graph, but in source code, it doesn't. In every time window, data share first time step graph. image image

Ripper346 commented 9 months ago

Hi, I am trying to interface for the first time with this library and create a suitable dataset myself. From the JSON of the dataset, I am assuming that the graph structure is always the same and the features change at each time period. Even though the description says, "The graph indicates how many people moved from one region to the other each day", the regions in England are always the same every day; what changes is the number of people from/to.