Awesome

Scalable Attributed-Graph Subspace Clustering (SASGC)

This repository provides Python code to reproduce experiments from the AAAI 2023 paper Scalable Attributed-Graph Subspace Clustering.

Run Experiments

Parameter List for `run.py`

Parameter	Type	Default	Description
`dataset`	string	`acm`	Name of the graph dataset (`acm`, `dblp`, `arxiv`, `pubmed` or `wiki`).
`power`	integer	`2`	First power to test.
`runs`	integer	`5`	Number of runs.

Best Propagation Orders

Dataset	Propagation order
`acm`	`2`
`dblp`	`2`
`arxiv`	`54`
`computers`	`67`
`wiki`	`4`
`pubmed`	`100`

Example

To run the model on computers for power p=67 and have the average execution time

python run.py --dataset=computers --power 67

Citation

If you use this code please do cite :

@inproceedings{fettal2023scalable,
  title={Scalable Attributed-Graph Subspace Clustering},
  author={Fettal, Chakib and Labiod, Lazhar and Nadif, Mohamed},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={37},
  year={2023}
}