The considerable growth of false information on social media has harmful consequences on all aspects of people's lives. There have been conducted various studies on designing automated fact-checking systems to promote the veracity and correctness of news and information to reduce the harmful effects of fake news and misinformation. Claim detection is regarded as the first step of developing fact-checking systems on which some studies have been done in a few languages. In the current paper, we aim to provide a corpus of Persian tweets and analyze them linguistically for annotation. Moreover, we develop a baseline claim detection model to evaluate the dataset. This study is framed as a classification task and enjoys transformer-based method to train a multi-label classifier to detect the different types of claims in Persian texts.

Keywords: Automatic Claim Detection, Persian Language Processing, Low-resource language, Multi Label Classification

Full-Text [DOCX 240 kb]   (22 Downloads)    

Add your comments about this article : Your username or Email:

---

---