mickeysclubhouse / covid-19-rumor-dataset

covid-19 rumor

File Structure

├── Data
│   ├── en_dup.csv
│	└── news
│	└── twitter
├── Data Analysis
│   └── PowerLaw Analysis.pdf
├── Data Collecting
│   └── snopes.py
├── LICENSE
└── README.md

Data Collecting

• snopes.py by Tianqi
  • It is used to collect data from website www.snopes.com and qc.wa.news.cn (departed)

Data Analysis

• PowerLaw Analysis includes the statistical analysis of rumor popularity data.
• Pre-processed data for deep learning tasks: https://drive.google.com/drive/folders/1jpnyHw1_XWR7lZTSb86AYWFsNC0bNEwO?usp=sharing.
• Veracity analysis https://colab.research.google.com/drive/1qtt0X9a7I9vBZPoZ-PzyhDuF5L_G3thj?usp=sharing
• Sentiment analysis https://colab.research.google.com/drive/1CyGtXPSol_Ayt_WR7zsq60M2Q5kDUofd?usp=sharing
• Stance analysis https://colab.research.google.com/drive/1ItGmEvumyOesXeB2dV7w5cnjRW9fTAKL?usp=sharing
• VAE model source https://www.researchgate.net/profile/Paul-Bogdan/publication/341128443_VRoC_Variational_Autoencoder-aided_Multi-task_Rumor_Classifier_Based_on_Text/links/5f1467fe92851c1eff1e647e/VRoC-Variational-Autoencoder-aided-Multi-task-Rumor-Classifier-Based-on-Text.pdf

Data

• news

  • news.csv (4129) and subfolder of each news
  • The number of subfolder records: 3936

• twitter

  • Twitter.csv (2705) and subfolder of each twitter
  • The number of subfolder records: 1383

• en_dup.csv

  • Unprocessed data with both news and twitter records.
  • The number of records: 7179 (with duplication).
  • Part of data are collected manually by keywords searching from sources such as twitter.com.
  • Data from www.snopes.com and qc.wa.news.cn are collected by 'snopes.py'.

Acknowledgement

• We thank Tianqi, Wenshuo, Jianni, Xiaofeng, and Hanlong for rumor data collection and labeling.

Cite Us

Cheng, Mingxi, et al. "A COVID-19 Rumor Dataset." Frontiers in Psychology 12 (2021): 1566.
@article{cheng2021covid,
title={A COVID-19 Rumor Dataset},
author={Cheng, Mingxi and Wang, Songli and Yan, Xiaofeng and Yang, Tianqi and Wang, Wenshuo and Huang, Zehao and Xiao, Xiongye and Nazarian, Shahin and Bogdan, Paul},
journal={Frontiers in Psychology},
volume={12},
pages={1566},
year={2021},
publisher={Frontiers}. }.
Link to paper: https://www.frontiersin.org/articles/10.3389/fpsyg.2021.644801/full

covid-19-rumor-dataset's Issues

dataset label(U, T, F) Meaning

Hello, I found 3 labels (F, T, U) in the data set. I want to know what they mean.

I guessed, F is no rumor, T is rumor, U is not sure whether it is a rumor

Thank you very much for your answers

Dataset sentiment label distributions

Hello, I found 4 sentiment labels (1-4) in the data set(data_sentiment.csv). However since the article(https://www.frontiersin.org/articles/10.3389/fpsyg.2021.644801/full#h8), it should be 5 labels included.

So I have 2 questions:

1. What these 4 labels represent for?
2. Why there is some label missing? and the distribution looks different with the article.

Thank you very much for your answers. It would be very helpful if I get the answers.