A Topic-aware Comparable Corpus of Chinese Variations
This provides a new dataset for researchers studying language variations, but it is incremental as it applies existing corpus construction methods to new data.
The study constructed a topic-aware comparable corpus of Mainland Chinese Mandarin and Taiwanese Mandarin from social media platforms (Sina Weibo and Dcard), addressing the gap in modern language resources by creating a regularly updated dataset.
This study aims to fill the gap by constructing a topic-aware comparable corpus of Mainland Chinese Mandarin and Taiwanese Mandarin from the social media in Mainland China and Taiwan, respectively. Using Dcard for Taiwanese Mandarin and Sina Weibo for Mainland Chinese, we create a comparable corpus that updates regularly and reflects modern language use on social media.