COVID-19 Rumor Detector
Brief Rumors Inference Model based on BERT Details- Obtained the original sample data of Dingxiangyuan (a medical knowledge sharing website) rumors and joint anti-rumor platform on the Internet with Python crawler.
- Segmented the corpus with Jieba, calculated word vectors and text relevance using SBERT, and applied BIRCH clustering and high frequency words analysis.
- Conducted fine-tuning of BERT models to realize semantic inferences (also NLI).