For details of the task, see FakeNewsChallenge.org
The data provided is (headline, body, stance) instances, where stance is one of {unrelated, discuss, agree, disagree}. The dataset is provided as two CSVs:
This file contains the body text of articles (the articleBody column) with corresponding IDs (Body ID)
This file contains the labeled stances (the Stance column) for pairs of article headlines (Headline) and article bodies (Body ID, referring to entries in train_bodies.csv).
The distribution of Stance classes in train_stances.csv is as follows:
| rows | unrelated | discuss | agree | disagree |
|---|---|---|---|---|
| 49972 | 0.73131 | 0.17828 | 0.0736012 | 0.0168094 |
Credits:
- Edward Misback
- Craig Pfeifer