SemEval 2019 Task 4 - Hyperpartisan News Detection
Authors/Creators
- 1. Bauhaus-Universität Weimar
- 2. Leipzig University
- 3. Factmata Ltd.
Description
Trial dataset for the SemEval 2019 Task 4: Hyperpartisan News Detection.
The dataset contains ~200.000 articles: ~100.000 hyperpartisan and ~100.000 least biased. 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. Some text is duplicated. Some tags are not cleaned yet. Also, some articles may be contained several times when they are published by several publishers. These errors will be fixed stepwise for the final data.
The final article data (within the <article> text) will contain only the following tags: <p>, <q>, <a>. The <a>-tags will have an attribute "type" that is either "external" (in which case the "href"-attribute contains the URI as usual) or "internal" (in which case the "href"-attribute is empty or left out). We removed all URLs that link to the same domain as the article ("internal") to avoid biasing classifiers towards utilizing the domain of the article. You will find that this tag-cleaning already succeeded for the vast majority of the articles in the current dataset, but unfortunately not all.
Files
Restricted
The record is publicly accessible, but files are restricted. Log in to check if you have access.
Request access
If you would like to request access to these files, please fill out the form below.
You need to satisfy these conditions in order for this request to be accepted:
Access is restricted to participants and organizers of the challenge for now. The data will be publicly available after the evaluation period.
You are currently not logged in. Do you have an account? Log in here
Additional details
Related works
- Is referenced by
- https://pan.webis.de/semeval19/semeval19-web/ (URL)