- Migyeong Yang, Chaewon Park, Jiwon Kang, Daeun Lee, Daejin Choi, Jinyoung Han, Fighting against Fake News on Newly-Emerging Crisis: A Case Study of COVID-19, ACM Web Conference 2024 (WWW '24 Companion).
This folder includes the whole dataset that we collected.
- COVID-19 Fake News Claim
- Non-COVID-19 Fake News Claim
- Crawled from Politifact and Snopes
- Topic Keyword
- Propagation Data
- Crawled from YouTube with GoogleAPI
- [train_ex1.pickle] Data used for training in Experiment 1, containing non-COVID-19 news claims.
- [train_ex2.pickle] Data used for training in Experiment 2, containing non-COVID-19 news claims and COVID-19 news claims published before April 1, 2020.
- [test.pickle] Data used for testing in Experiment 1 and Experiment 2, containing COVID-19 news claims published after April 1, 2020.
This folder contains only the data necessary for model training.
- [train_ex1.pkl] A version extracted from './dataset/train_ex1.pickle', including news claims (title), ner_onehot, titles and descriptions of two YouTube videos, and labels.
- [train_ex2.pkl] A version extracted from './dataset/train_ex2.pickle', including news claims (title), ner_onehot, titles and descriptions of two YouTube videos, and labels.
- [test.pkl] A version extracted from './dataset/test.pickle', including news claims (title), ner_onehot, titles and descriptions of two YouTube videos, and labels.
- [train_ex1.npy / train_ex2.npy / test.npy] These are numpy-formatted adjacency matrices required for training and testing.
This folder consists of Python scripts for the model.
Python 3.7 & Pytorch 1.10.1
pip install -r requirements.txt
For Exp. 1 (Without COVID-19 Data)
python3 ./model/main-ex1.py
For Exp. 2 (With a few COVID-19 Data)
python3 ./model/main-ex2.py
BIBTEX
@inproceedings{yang2024fighting,
title={Fighting against Fake News on Newly-Emerging Crisis: A Case Study of COVID-19},
author={Yang, Migyeong and Park, Chaewon and Kang, Jiwon and Lee, Daeun and Choi, Daejin and Han, Jinyoung},
booktitle={Companion Proceedings of the ACM on Web Conference 2024},
pages={718--721},
year={2024}
}