TUMCC is the first Chinese corpus in the jargon identification field.
28,749 sentences, including 804,971 characters, from 19,821 Telegram users of 12 Telegram groups were collected when we built TUMCC.
We had finished data screening and word segmentation before we released this corpus. So it might be easier for you to use.
After cleaning, TUMCC contains 3,863 sentences (100,000 characters) from 3,139 Telegram users.
TUMCC-clean.txt contains the corpus after our cleaning. You can use it directly in your research.
TUMCC-raw.7z contains raw information we collected from Telegram. You can do text cleaning to get more valid data and valuable information.
For more details about the target Telegram group sources for data extraction, please refer to the paper Identification of Chinese Dark Jargons in Telegram Underground Markets Using Context-Oriented and Linguistic Features (Information Processing and Management, 2022).
Thanks for your interest in our dataset, please feel free to leave a ⭐️ or cite us through:
@article{hou2022identification,
title={Identification of Chinese dark jargons in Telegram underground markets using context-oriented and linguistic features},
author={Hou, Yiwei and Wang, Hailin and Wang, Haizhou},
journal={Information Processing \& Management},
volume={59},
number={5},
pages={103033,1--20},
year={2022},
publisher={Elsevier}
}