https://github.com/thunlp/mmdw
Max-margin DeepWalk
https://github.com/thunlp/mmdw
network-embedding
Last synced: about 1 year ago
JSON representation
Max-margin DeepWalk
- Host: GitHub
- URL: https://github.com/thunlp/mmdw
- Owner: thunlp
- License: mit
- Created: 2016-04-18T02:42:51.000Z (about 10 years ago)
- Default Branch: master
- Last Pushed: 2017-05-09T09:07:36.000Z (about 9 years ago)
- Last Synced: 2025-04-03T20:23:35.445Z (about 1 year ago)
- Topics: network-embedding
- Language: Java
- Size: 1.84 MB
- Stars: 71
- Watchers: 12
- Forks: 38
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# MMDW
Max-margin DeepWalk
Code of IJCAI2016: "Max-Margin DeepWalk: Discriminative Learning of Network Representation"
## Datasets
We evaluate MMDW with three datasets, including Cora, Citeseer and Wiki.
* data/Edgelist/*_edgelist.txt: original edgelist file of network *.
* data/Category/*_category.txt: the category list of vertices.
* data/Net/*_net.txt: the matrix M of obtained from transition matrix A of a network.
* data/vector/: the folder to save learnt vectors of vertices.
* data/svm_model/: the folder to save trained svm classifiers.
* data/Bias/: the folder to save calculated biasVectors.
* data/result/: the folder to classification results.
## Run
Run the following command to learn max-margin DeepWalk:
java -jar mmdw.jar "dataset" "data_folder" "order_of_alphaBias"
Here is an example:
java -jar mmdw.jar Cora data/ -3
## Cite
If you use the code, please cite this paper:
_Cunchao Tu, Weicheng Zhang, Zhiyuan Liu, Maosong Sun. Max-Margin DeepWalk: Discriminative Learning of Network Representation. International Joint Conference on Artificial Intelligence (IJCAI 2016)._
For more related works on network representation learning, please refer to my [homepage](http://thunlp.org/~tcc/).