THUOCL

Why we included this project

Chinese word segmentation is one of those tasks where the dictionary you feed the segmenter decides the outcome, and THUOCL gives you a set of hand-curated lists built for exactly that. The eleven domains cover IT, finance, idioms, place names, historical figures, poetry, medical, food, law, vehicles, and animals, and every entry carries a document-frequency value, so you can filter by how common a term is across the corpus it was counted on. That matters if you want a vocabulary tuned to your own text instead of merging everything blindly. Segmenters like THULAC and Jieba can take these plain-text lists to keep proper nouns and domain jargon from being chopped into fragments. It is a data resource, not an application, so plan on feeding the tab-separated files into your own tokenizer rather than running anything out of the box.