The source code of article titled "The Effects of Media Bias on News Recommendations"
Hyperpartisan News Detection Dataset is a media bias detection dataset which classifies the news articles into bias news articles and unbias news articles. More details: Data for PAN at SemEval 2019 Task 4: Hyperpartisan News Detection
MIcrosft News Dataset (MIND) is a large scale English-based news recommendation dataset. More details: MIND: MIcrosoft News Dataset
The structure of the code is as follows:
- bias_detector: the source code for detecting media bias
- ms_recommender_original: NPA, NAML, LSTUR and NRMS are implemented using the source code from Microsoft Recommenders repository.
- news_recommendation: FIM, PLM-empowered are re-implemented under this folder.
- observer: keep source code for generating user groups; present analysis results based on the inference results generated by news recommendation algorithms under different user groups described in our article.