Dataset Open Access
SemEval 2019 Task 4 - Hyperpartisan News Detection
Johannes Kiesel;
Martin Potthast;
Maria Mestre;
Rishabh Shukla;
Benno Stein;
David Corney;
Emmanuel Vincent;
Payam Adineh
Third trial dataset for the SemEval 2019 Task 4: Hyperpartisan News Detection.
The dataset contains 1 million articles. It is split in training (200,000 left, 400,000 least, 200,000 right) and validation (50,000 left, 100,000 least, 50,000 right), where no publisher that occurs in the training set also occurs in the validation set. All articles are labeled by the overall bias of the publisher as provided by BuzzFeed journalists or MediaBiasFactCheck.com.
The trial data is not fully cleaned. Due to some encoding error, some characters are replaced by question marks. However, all files are already fully compatible with the XML schema files.
| Name | Size | |
|---|---|---|
|
article.xsd
md5:31e3fb439c98b18cd74a7d936a65b218 |
2.1 kB | Download |
|
articles-training-20180831.xml.zip
md5:c3e85da69f0ec76d30a2c1a0b22d3150 |
1.4 GB | Download |
|
articles-validation-20180831.xml.zip
md5:5dd17f5043f130407cf599d585ba4ca9 |
547.9 MB | Download |
|
ground-truth-training-20180831.xml.zip
md5:7ea315edde4f500b554571f388a2fa46 |
30.0 MB | Download |
|
ground-truth-validation-20180831.xml.zip
md5:50fdbd01f9eef4902a0e6fe93a360577 |
7.0 MB | Download |
|
ground-truth.xsd
md5:81dd0e153d6f78ca10a5599da6aac66e |
1.6 kB | Download |
| All versions | This version | |
|---|---|---|
| Views | 21,135 | 8,932 |
| Downloads | 24,031 | 1,952 |
| Data volume | 10.6 TB | 982.7 GB |
| Unique views | 17,168 | 8,432 |
| Unique downloads | 5,745 | 550 |
- Publication date:
- September 3, 2018
- DOI:
-
- Keyword(s):
- Hyperpartisan news SemEval SemEval 2019 SemEval 2019 Task 4 Biased news News bias Hyperpartisan Hyperpartisanship
- Meeting:
- International Workshop on Semantic Evaluation 2019 (SemEval-2019) (Session Task 4)
- Related identifiers:
- Referenced by
- Communities:
- License (for files):
- Creative Commons Attribution 4.0 International