Skip to content
master
Switch branches/tags
Go to file
Code

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

README.md

HGATRD

The implementation of our IJCNN 2020 paper "Heterogeneous Graph Attention Networks for Early Detection of Rumors on Twitter"

Requirements

python 3.6.6
numpy==1.17.2
scipy==1.3.1
pytorch==1.1.0
scikit-learn==0.21.3

How to use

Dataset

The main directory contains the directories of two Twitter datasets: twitter15 and twitter16. In each directory, there are:
ind.twitter1*.label files: the labels of all nodes in the tweet-word subgraph.
ind.twitter1*.features_index files: the index of words that nodes in the tweet-word subgraph contained.
ind.twitter1*.adj files: the weight adjacency matrix of the tweet-word subgraph.
ind.twitter1*.train files: the index of the training set in the tweet-word subgraph.
ind.twitter1*.dev files: the index of the validation set in the tweet-word subgraph.
ind.twitter1*.test files: the index of the testing set in the tweet-word subgraph.
ind.twitter1*.user.tweet.adj files: the weight adjacency matrix of the tweet-user subgraph.
ind.twitter1*.user.tweet.train files: the index of the training set in the tweet-user subgraph.
ind.twitter1*.user.tweet.dev files: the index of the validation set in the tweet-user subgraph.
ind.twitter1*.user.tweet.test files: the index of the testing set in the tweet-user subgraph.
ind.twitter1*.user.tweet.adj.tc* files: the weight adjacency matrix of the tweet-user subgraph which simulated the task of early detection by controlling the received tweets.
ind.twitter1*.user.tweet.adj.et* files: the weight adjacency matrix of the tweet-user subgraph which simulated the task of early detection by controlling the elapsed time.
These datasets are preprocessed according to our requirement and original datasets can be available at here

Training & Testing

sh run.sh 0 twitter15[twitter16]

Citation

If you find the code is useful for your research, please cite this paper:

@article{huang2020heterogeneous,
  title={Heterogeneous Graph Attention Networks for Early Detection of Rumors on Twitter},
  author={Huang, Qi and Yu, Junshuai and Wu, Jia and Wang, Bin},
  journal={arXiv preprint arXiv:2006.05866},
  year={2020}
}

About

No description, website, or topics provided.

Resources

Releases

No releases published

Packages

No packages published