Graph signal classification

This is a graph signal classification dataset.

Data is described in our publication, available online at:

https://proceedings.mlr.press/v238/xu24c/xu24c.pdf

The dataset consists of 3 graph signal classification tasks, which are labeled by the folders partly_cloudy, synthetic, traffic

Partly cloudy data is originally from:

Richardson, H., Lisandrelli, G., Riobueno-Naylor, A., and Saxe, R. (2018). Development of thesocial brain from age three to twelve years. Nature communications, 9(1):1027.

We used the version of the data presented in:

Rieck, B., Yates, T., Bock, C., Borgwardt, K., Wolf, G., Turk-Browne, N., and Krishnaswamy, S.(2020). Uncovering the topology of time-varying fmri data using cubical persistence. Advances inneural information processing systems, 33:6900–6912.

The dataset consists of fMRI signals, and the label corresponds to the emotional state of the film.

Synthetic data is generated according to a procedure described in our paper and is provided for convenience. 

The traffic dataset is from the Caltrans Performance Measurement System (PeMS), and is preprocessed following:

Guo, S., Lin, Y., Feng, N., Song, C., and Wan, H. (2019). Attention based spatial-temporal graph convolutional networks for traffic flow forecasting. In Proceedings of the AAAI conference on artificial intelligence, volume 33 No. 01, pages 922–929.

The graph signal labels are the time of the measurement at different granularities.