junwu6 / marine

Scalable Manifold-Regularized Attributed Network Embedding via Maximum Mean Discrepancy

MARINE (MAnifold-RegularIzed Network Embedding)

An implementation for "Scalable Manifold-Regularized Attributed Network Embedding via Maximum Mean Discrepancy" (CIKM'19). [Paper]

Environment Requirements

The code has been tested under Python 3.6.5. The required packages are as follows:

• tensorflow == 1.12.0
• numpy == 1.15.4
• scipy == 1.1.0
• sklearn == 0.20.0
• networkx == 2.3

Data sets

We used four data sets in our experiments: Cora, Citeseer, Pubmed and Wiki.

Run the Codes

python main.py

Acknowledgement

This is the latest source code of MARINE for CIKM2019. If you find that it is helpful for your research, please consider to cite our paper:

@inproceedings{wu2019scalable,
  title={Scalable Manifold-Regularized Attributed Network Embedding via Maximum Mean Discrepancy},
  author={Wu, Jun and He, Jingrui},
  booktitle={Proceedings of the 28th ACM International Conference on Information and Knowledge Management},
  pages={2101--2104},
  year={2019},
  organization={ACM}
}

marine's Issues

Classification accuracy of PubMed dataset

Hello, I run your code on the PubMed dataset. When I run the training on 70% of the nodes, I set it according to the parameters in the paper, and other parameters use the default values. The classification accuracy is about 0.58, which is quite different from that in your paper. The parameters I set may not be optimal. For the PubMed dataset, can you give us the specific parameter settings? We look forward to your reply!

can you upload cluster files and link prediction files?

Hello, your work is very interesting, can you upload cluster files and link prediction files?Looking forward to your reply！