TMIP_ED: A dataset of entity annotations in Tibetan medical classics

We proposes a Hybrid Adaptive Word Segmentation (HAWS) method based on a previously constructed Tibetan Medicine entity dictionary, to extract entity terms from ancient Tibetan medical texts and to build and release a Tibetan Medicine entity dataset (TMIP_ETD). Specifically, First, the 100 ancient books (333 volumes) included in the “Collection of Classical Tibetan Medicine Literature” were digitized, resulting in a text dataset of 13.8 million syllables.

Then, the HAWS method was used for tokenization, and potential entities with a frequency of at least 12 were extracted. After manual verification, a dataset of Tibetan medicine classics entity terms containing 34,601 entries was obtained. This Dataset providing data foundation for knowledge mining and digital inheritance of Tibetan medicine ancient classics.