// the find
liuhuanyong/DomainWordsDict
DomainWordsDict, Chinese words dict that contains more than 68 domains, which can be used as text classification、knowledge enhance task。涵盖68个领域、共计916万词的专业词典知识库，可用于文本分类、知识增强、领域词汇库扩充等自然语言处理应用。
A static collection of 9.16 million Chinese domain-specific words across 68 categories, each word tagged with a representativeness weight. Useful as a feature set for Chinese NLP tasks — text classification, entity recognition, domain detection — particularly when you don't have labeled training data and want a fast keyword-matching baseline.
The weight column is the real value here: words are ranked by how well they discriminate their domain, so you can tune a threshold rather than dump everything into your pipeline. Coverage is unusually wide — 68 domains from metallurgy to online gaming to classical poetry, which beats most Chinese NLP starter packs. The format is dead simple: tab-separated, one word per line, no tooling required to use it. For zero-shot domain classification via keyword counting, this will outperform a generic tokenizer on specialized text.
Last updated in 2021 and showing it — the 'mobile/digital' category still lists Nokia-era brands, and anything touching internet culture or recent tech will have gaps. Some categories have serious quality problems: '新番动漫' (anime) and '人物名称' (people names) are filled with garbled Unicode characters that look like encoding artifacts, not actual words. No tooling, no loader, no Python package — it's a data dump, so you own the integration work entirely. The domain taxonomy is also fixed; if your domain doesn't map to one of the 68, you're stuck building your own weights from scratch.