Awesome

#PRRE

This is a python implementation of my CIKM 2018 paper "PRRE: Personalized Relation Ranking Embedding for Attributed Networks", which considers the partial correlation between node attributes and network topology.

This is the original version of PRRE, the advanced version will be released later

Requirements

• networkx
• numpy
• scipy
• scikit-learn

All required packages are defined in requirements.txt. To install all requirement, just use the following commands:

pip install -r requirements.txt

Basic Usage

Input Data

Each dataset contains 3 files: edgelist, feature for training and label for evaluation.

1. dataset.edgelist: each line contains two connected nodes.
node_1 node_2
node_2 node_3
...

2. dataset.feature: this file has n lines.
feature_1 feature_2 ... feature_n
feature_1 feature_2 ... feature_n
...

3. dataset.label: each line represents a node and its class label.
node_1 label_1
node_2 label_2
...

Run

To run PRRE, just execute the following command for node classification task and link prediction with different 'task' parameter:

python prre.py

For data visualization task, Embedding Projector is recommended to deal with the embedding result.

Cite

@inproceedings{zhou2018prre, title={PRRE: Personalized Relation Ranking Embedding for Attributed Networks}, author={Zhou, Sheng and Yang, Hongxia and Wang, Xin and Bu, Jiajun and Ester, Martin and Yu, Pinggang and Zhang, Jianwei and Wang, Can}, booktitle={Proceedings of the 27th ACM International Conference on Information and Knowledge Management}, pages={823--832}, year={2018}, organization={ACM} }